FCA Guidelines on

Risk Assessment of non-listed substances (NLS) and

non-intentionally added substances (NIAS) under the

requirements of Article 3 of the Framework

Regulation (EC) 1935/2004

Version 1.0

October 2016

1

This document aims to give guidance to manufacturers or downstream users of food contact substances in regard to the Risk Assessment of non-listed substances (NLS) and non-intentionally added substances (NIAS) in order to fulfill the requirements of Article 3 of the Framework Regulation (EC) 1935/2004. This document is provided for general guidance information purposes only. This is a living document which will be updated when needed.

2

TABLE OF CONTENTS

EXECUTIVE SUMMARY

3

1. INTRODUCTION

4

2. DEFINITIONS

7

3. TOXICOLOGICAL ASSESSMENT

10

3.1 Determination Tolerable Daily Intake (TDI) based on specific toxicological studies 11 3.2 Uncertainty factors 13 3.3 The Use of the DNEL (REACH) for Risk Assessment in Food Contact 18 3.4 Read across 19 3.5 Determination Threshold of Toxicological Concern (TTC) 23 3.6 10ppb- approach / 50ppb- approach 28 3.7 Nano 30 3.8 Endocrine disruptors 31 3.9 Bioaccumulation / bioavailability in the human body 31

4. EXPOSURE ASSESSMENT

32

4.1 Tiered approach 32

4.2 Migration Estimation 33

4.3 Food consumption and packaging use data 37

4.4 Derivation Estimated Daily Intake number – three steps approach 37

4.5 European consumer exposure assessment tools 40

5. RISK CHARACTERISATION OR RISK ASSESSMENT

47

REFERENCES

48

3

EXECUTIVE SUMMARY

The present guidelines were developed by FCA – “Food Contact Additives” Sector Group of Cefic –

the European Chemical Industry Council. It aims at providing guidance on risk assessment principles

to manufacturers and downstream users of substances used in food contact materials. The

guidelines provide an overview of methods and approaches for risk assessing so-called “non-listed

substances” and “non-intentionally added substances” in order to fulfil the requirements of Article 3

of the Framework Regulation (EC) 1935/2004. This document therefore applies to substances

exempted from authorization or for applications where no specific legislation exists.

Under chapters 1 and 2, the scope of the guidelines and relevant definitions are provided. Chapter 3

illustrates different internationally recognized scientific principles on toxicological assessment of

substances used in various food contact materials applications. Chapter 4 highlights some tiered

approaches for estimating the exposure of a given substance to the consumer. While chapter 5

indicates the principles for finalizing the risk assessment based on the findings gathered in the two

previous chapters.

The present Guidelines are to be regarded as a living document, which will be regularly updated,

when necessary in order to take into account latest scientific developments in this area.

4

1. INTRODUCTION This document aims to give guidance to manufacturers or downstream users of food contact

substances in regard to the risk assessment of non-listed substances (NLS) and non-intentionally

added substances (NIAS) in order to fulfill the requirements of Article 3 of the Regulation (EC) No

1935/2004 on materials and articles intended to come into contact with food (Abbrev: “Framework

Regulation No 1935/2004 on Food Contact Materials”). However, this document is provided for

general guidance information purpose only. The use of any of these risk assessment methodologies

is at the users own risk, and it is recommended to first seek scientific or legal advice. The

procedures mentioned in this guidance do not constitute legal advice or opinions of any kind, or any

advertising or solicitation. The authors of this guidance will not be liable for any damages, losses or

causes of action of any nature arising from any use of risk assessment methodologies.

This is a living document which will be updated when and if needed.

Taking into account the actual state of the art, some applications like the plastic food contact

materials and articles are regulated by a specific measure - the Commission implementing

Regulation No 10/2011 on plastic materials and articles intended to come into contact with food

(Abbrev:“ Commission implementing regulation No 10/2011 on plastic materials and articles”) - while

most of the other applications, e.g. coatings, printings, paper and board, adhesives, are not subject

to harmonized EU legislation. National provisions might exist in one or more countries of the EU.

Plastic food contact materials and articles

In the Commission implementing regulation No 10/2011 on plastic materials and articles, there is a

mandatory listing of monomers, starting substances, macromolecules obtained from microbial

fermentation and additives. There are also substances for which an exemption of listing exists.

These substances need to be assessed in accordance with internationally recognised scientific

principles on risk assessment as described in article 19 of the Regulation. These substances can be

classified as non-listed substances (NLS) and non-intentionally added substances (NIAS) and may

be present in small quantities with a potential or not to migrate into the food or food simulants.

Included in those both categories, there are migrating by-products, decomposition products or other

residues, but also intentionally added substances such as aids to polymerisation (AP) like e.g.

catalysts, initiators and “polymer production aids” (PPA), solvents, emulsifiers, etc.

Non-plastic food contact materials and articles

In the absence of EU harmonized specific measures for the so-called “non-plastics”, a similar

principle, as expressed in Article 19 of Commission implementing Regulation (EU) No 10/2011 on

plastic materials and articles, is needed for those substances for which no mandatory authorisation

is requested. These self-assessments based on internationally recognized principles will serve the

requirements of Article 3 of the Framework Regulation No 1935/2004 on Food Contact Materials

which stipulates that all migrating substances in food contact materials have to be safe and do

not bring unacceptable changes to the food.

Based on the principle of Article 3 of the Framework Regulation No 1935/2004 on Food Contact

Materials and the Responsible Care principle, all migrating substances (listed and non-listed as well

as NIAS) should be present in quantities as low as reasonably and technically possible. The

principles of risk assessment should never be used to whitewash avoidable migration and residues

5

in FCM. Appropriate quality measures should always be in place to produce a food contact material

as suitable as possible in order to minimize migration.

A risk assessment consists of three components: hazard identification and characterisation,

exposure assessment, followed by the risk assessment itself. Hazard is the potential of something to

cause harm. Hazard typically refers to the intrinsic properties of a chemical, such as toxicity, while

exposure addresses the likelihood and degree to which a human or environmental receptor will be

exposed to the intrinsic hazards of a chemical. Risk is the likelihood of harm occurring.

Captured into a simple formula, this would read: Hazard x exposure potential = risk.

Risk assessment puts hazard and exposure together in an attempt to understand the “real world

danger” posed by a chemical based on its intrinsic hazards in the light of anticipated exposure.

Risk assessment certainly requires the rigorous and objective analysis of data. However, there may

be weaknesses or gaps in the data that can only be addressed by applying professional judgment.

The various assumptions and uncertainties carried over from the hazard characterisation and

exposure assessments also affect the risk assessment. Extrapolating from that work to reach

probabilistic conclusions necessarily creates additional uncertainties.

For those non-listed substances which do not need to be authorized or which are only present as

residues, the following methodology could be used:

• Verify if the substance is authorized by international recommendations or at national level; or

• Do a risk assessment on the basis of internationally recognised scientific principles according

to Article 19 of Commission implementing Regulation (EU) No 10/2011 on plastic materials

and articles resp. Article 3 of the Framework Regulation No 1935/2004 on Food Contact

Materials.

What does this paper intend to deliver?

This Guideline aims to provide hands on support for companies which are in the process of

assessing their products and/substances by giving an overview of methods how to risk assess non-

listed substances and non intentionally added substances in food contact materials in order to meet

Art 3 of the Framework Regulation No 1935/2004 on Food Contact Materials.The approach is similar

to Article 19 of Commission implementing Regulation (EU) No 10/2011 on plastic materials and

articles. Under the latter’s provisions, a number of substances present in food contact plastics are

exempt from the requirement to be included in the Community positive list (Union list) according to

Article 6 of Commission implementing Regulation (EU) No 10/2011. The substances exempted of

positive listing include: solvents, colourants, polymer production aids (PPA’s), aids to polymerisation

(AP’s), oligomers and non-intentionally-added-substances (NIAS). NIAS include substances such as

impurities, contaminants, reaction - decomposition- or degradation- products. It should therefore be

made clear that this guidance document applies to substances exempted from authorisation or for

applications where no specific legislation exists.

Until now, the term NIAS has not been used in European legislation for non-plastic FCMs, with the

exception of the Dutch Commodities Act regulation on packagings and consumer articles coming

6

into contact with foodstuffs in the revision dated March 2014. Since NIAS do not only occur in

plastics but may also be present in non-plastic FCMs such as paper/board, coatings, metals, cork,

etc, the term NIAS used in this guideline is used in accordance with with Article 3 of Regulation EU

No 10/2011 where NIAS are defined as: non-intentionally added substance means an “impurity in

the substances used or a reaction intermediate formed during the production process or a

decomposition or reaction product”. Depending on the physical/chemical parameters and the

chemical composition of FCMs, and on the nature of the food, FCMs and articles may transfer their

constituents (both intentionally added substances – IAS - and NIAS) to foods. This mass transfer

phenomenon is called migration. Migration may lead to exposure to certain chemicals, which might

cause or might not cause a risk for human health and so it must be evaluated and controlled.

Furthermore, migration which brings about an unacceptable change in the composition of the food or

brings about deterioration in the organoleptic properties of the food must be avoided.

Regarding the risk assessment, in most cases, only migrants up to a molecular weight (MW) of 1000

Daltons (Da) have to be considered. This threshold of 1000 Da is important as the European Food

Safety Authority (EFSA) has conventionally assumed in its assessments of plastics starting materials

that above this molecular weight, substances are not absorbed by the body and therefore may be

excluded from any calculations of migration and exposure (EFSA, 2008).

7

2. DEFINITIONS

Absolute Barrier: A material at a given thickness which excludes any permeation of potential

migrants from outside into the packed product under any foreseeable contact conditions.

Examples include:

- Glass of any thickness (not: SiOx layers).

- Metal cans and lids.

- Aluminium foils at thickness when pinholes or other damages can be excluded.

CMR Substance: A substance listed as Carcinogenic, Mutagenic or toxic to Reproduction

Category 1A, 1B or 2 in CLP Regulation 1272/2008 Annex VI table 3.1.

Estimated daily Intake: Amount of a substance that is daily absorbed by oral route by the

consumer through food contact articles.

Exposure: In the context of food contact, exposure refers to consumer exposure, and more

particularly to the quantity of a substance that a consumer is exposed to, by ingestion. It can be

further refined to refer to specific consumer groups, e.g. children or infants, or to a statistical

fraction of the consumer population, e.g. those consumers exposed to 95percentile level (which

means that only 5% of the population is considered as not being covered).

Functional Barrier: One or more layers of any material type which limits the transfer of relevant

substances into the packed product to either a) less than 10 ppb or b) below another level of

regulatory or safety concern. For example, in the context of the Plastics Regulation 10/2011,

Functional Barrier means a barrier consisting of one or more layers of any type of material which

ensures that the final material or article complies with Article 3 of Regulation (EC) No 1935/2004

and with the provisions of the regulation. Useful information on what might constitute a functional

barrier can be found in Section 5.2.7 of Technical Guidelines for Compliance Testing. The

functional barrier concept does not cover substances which are mutagenic, carcinogenic or toxic

to reproduction or to substances in nano form.

Hazard: Potential source of harm (adverse effect) or an intrinsic ability to cause harm

Migratable Substance: A chemical substance that is capable of transfer in detectable amounts

from the food contact material and/or article into the food.

Overall migration limit (OML): The maximum permitted amount of non-volatile substances

released from a material or article into food simulants.

Non-listed substance (NLS): An intentionally added substance which is exempted from the

authorisation process, meaning which is exempted from a positive listing. An example of

exempted substances is solvents.

Non-Intentionally added substances (NIAS): An impurity in the substances used or a reaction

intermediate formed during the production process or a decomposition or reaction product.

8

Repeated use article: an article intended to be used several times that comes into contact with

different portions of foods during its lifetime.

Risk: Risk is the chance or probability of harm (adverse effect) if exposed to a hazard. Risk is a

function of hazard and exposure

Risk Analysis: Reviews and communicates the results from the risk assessment and risk

management processes

Risk Assessment: A process to provide an understanding of the hazard posed in light of the

anticipated exposure

Risk Management: Decision-making/policymaking based on the results of risk assessment and

stakeholder input

Rubber: low shear modulus materials, either natural1 or synthetic, made up of carbonaceous

macromolecules, and characterised by long polymer chains arranged in a three-dimensional

flexible network held by chemical covalent cross-links. They present, at service temperature and

until their decomposition, elastic physical properties which allow the material to be substantially

deformed under stress and recover almost its original shape when the stress is removed. The

definition does not cover thermoplastic elastomers.

Rubber products: finished material and article constituted of rubber including thermoplastic

rubber as well as blends of rubber with plastics and other materials, which are intended to come

into contact with or are placed in contact with foodstuffs. A rubber product may be made almost

entirely of rubber, e.g.a glove, or it may contain components and reinforcement other than rubber,

as for an example in a rubber coated fabric, a tyre, a steel laminated bridge bearing and a rubber

hose fitted with a metallic coupling.2

Set-off: The phenomenon of the transfer of substances from outer layer of materials and articles

to the inner food contact layer through direct contact and not via diffusion through the material.

Set-off may occur, where there is a contact between the outside and inside of the material or

article during, for example, storage or transport. Such direct contact may occur when materials

are wound in reels or stacked in sheets or when articles such as trays and pots are nested inside

each other. Unlike migration under these conditions, set-off may occur in both materials and

articles with or without a functional barrier.

Simulant: A test medium used to represent a type of packed product when measuring the

migration of substances from the packaging. Food simulants are specified in Annex III of the

Commission Implementing Regulation No 10/2011.

Specific migration Limit (SML): The maximum permitted amount of a given substance released

from a material or article into food or food simulants.

1 For example, caoutchoucs which are naturally derived rubber from latex originating from the sap of trees.

2 Resolution of the Council of Europe AP (2004)4 on rubber products intended to come into contact with foodstuffs_

Version 1_10.06.2004

9

Thermoplastic elastomers TPE): Polymer or blend of polymers that does not require

vulcanisation or cross-linking during processing, yet has properties, at its service temperature,

similar to those of vulcanised rubber. These properties disappear at processing temperature, so

that further processing is possible, but return when the material is returned to its service

temperature. They are covered under the definition of plastics under Commission implementing

Regulation (EU) No 10/2011.

Worst Case Calculation: This assumes that all the migrants will transfer from the packaging

material into the packed product. To do it you need:

- Maximum concentration (MC) of substance in the material layer (in ppm).

- Maximum grammage per square metre (G) of material layer (or thickness and density) (in

g/m² ).

- Maximum surface area of packaging to weight of food packed (SV) (in dm²/ kg).

The formula to calculate maximum amount of substance that could migrate into the food is

MC x G x SV ÷ 100000 mg/kg (ppm)

10

3. TOXICOLOGICAL ASSESSMENT Toxicological assessment aims to identify the adverse toxicological effects that a substance could

cause (i.e. hazard identification) and secondly, to define the critical dose or exposure level of a

substance in the daily diet, below which the substance is not expected to pose a risk to human

health (i.e. dose response assessment or hazard characterisation). So the aim of this chapter is to

provide an overview on methodologies about how to define a safe dose of a given substance.

Most adverse effects for chemicals occur at a particular dose (Paracelsus: “dose makes the poison”).

Toxicological studies or alternative data will be applied to derive the daily dose which can, based on

conservative assumptions, be expected with reasonable certainty to be safe.

This critical dietary exposure level is often referred to as the Tolerable Daily Intake (TDI), generally

used for substances appearing in food but not intentionally added or the Acceptable Daily Intake

(ADI) for substances intentionally added to food, usually expressed in mg/person/day or mg/kg

bodyweight/day. The TDI concept is based on the assumption that a clear dose-repsonse

relationship with a threshold exists, whereas the threshold defines the point of exposure below which

no adverse effect is observable. The TDI is traditionally derived from a NOAEL (no-adverse-effect-

level) using animal studies.

As an alternative, EFSA proposes the use of the benchmark dose (BMD) and the Scientific

Committee concludes “that the BMD approach is a scientifically more advanced method to the

NOAEL approach for deriving a Reference Point, since it makes extended use of available dose-

response data and it provides a quantification of the uncertainties in the dose-response data”.3

Based on the TDI and the current4 European default assumption that a 60 kg person consumes a

kilogram of food per day, a self-derived Specific Migration Limit (SML) for the substance can be

calculated using the following formula:

Self-derived SML (mg/kg food) = 60 (kg body weight) * TDI (mg/kg body weight/day) / 1 kg

food/day

For some adverse effects a clear dose-response relationship cannot be defined or does not exist.

The derivation of a safe dose is therefore impossible and other concepts need to apply.

Genotoxic, mutagens and carcinogens are examples of substances where no clear dose-response

relationship may exist. For their mode of action, it is traditionally assumed that already one

interaction event between a substance molecule and a DNA molecule could theoretically lead to an

adverse effect, so that a no-threshold-mechanism is assumed5. Generally, the aim is to strictly avoid

the presence of genotoxic, mutagens and non-threshold carcinogens in food contact materials.

3Guidance of the Scientific Committee on a request from EFSA on the use of the benchmark dose approach in

risk assessment. The EFSA Journal (2009) 1150, 1-72 4 This is the situation at the current point in time. EFSA is in the process of revising their note for guidance,

which might have an impact on this current assumption 5 There is on-going scientific debate about this hypothesis and the consensus may change in the near future,

but this has to be discussed elsewhere and the current guidance document will build on the traditional hypothesis and risk assessment methods.

11

However, this may not always be possible, especially for NIAS. A safety assessment for such cases

would follow the internationally accepted scientific principles of linear low dose extrapolation, the

Margin of Exposure (MOE) approach or Derived Minimal Effect Levels approach. In the case of

food contact materials applied in Europe, the MOE approach is preferable, as it has been reviewed

and recommended by the EFSA Scientific Committee11.

Regulatory agencies in the United States, the Food and Drug Administration (FDA) and the

European Union (EU) use a tiered approach based on the “dose makes the poison” principle to

regulate substances that e.g. migrate from food packaging and processing equipment to food.

Toxicological data may not be required when the exposure is extremely low. Under U.S. FDA

guidance, substances with an exposure below the Threshold of Regulation of 1.5 µg/person/day and

no concern of genotoxicity do not require specific toxicological data. For Europe, under the

provisions of the Regulation No 10/2011, substances that have not been evaluated and authorized

and are not classified as carcinogenic, mutagenic or reprotoxic, can be used in plastics layers behind

a functional barrier if they do not migrate at a detection limit of 10 g/kg food. This “no-migration”

concept for non-CMR substances has been adopted under the CEPE Code of Practice for non-listed

substances in direct food contact coatings.

The first step of a safety assessment is always the search for toxicity data on the substance.

Subsequently, there are basically two approaches to determine the dietary exposure thresholds for

substances:

- The determination of a tolerable daily intake (TDI), based on toxicological studies performed

on the substance or a structurally similar substance (read across); or

- If no substance specific data are available, use the Threshold of Toxicological Concern (TTC)

concept as a basis

Substances being suspected or known genotoxins and/or carcinogens require specific risk

assessment methodology which shall not be discussed here. For guidance, please refer to the MOE

approach.

3.1 Determination Tolerable Daily Intake (TDI) based on specific toxicological studies

DATA AVAILABILITY

1) The first step is to search for all toxicological data available for the substance (including its

impurities) or for similar substances / category of substances and to define the degree of purity

needed for food contact materials. The focus should be especially on:

Mutagenicity tests (in-vitro and in-vivo).

Repeated dose studies (28-d, 90-d oral or chronic/dermal/inhalative toxicity study).

Studies on absorption, distribution, metabolism and excretion (ADME).

Carcinogenicity studies (oral, dermal, inhalative).

Studies on reproduction and developmental toxicity (oral, dermal, inhalative).

12

Studies should be performed according to OECD test methods or international standardised test

guidelines; if not available, the test methods described in the study should be judged against these

standardised methods.

The following data sources6 may be used (Peer reviewed data sources are marked in bold below)

In-house data / owned data from other (regulatory) sources (incl suppliers).

REACH data:

- REACH registration dossiers on ECHA website (reference as given in the document)

- ECHA/Member States peer-reviewed information on toxicological data for

REACH/CLP/Biocides7.

ECHA Final decisions on compliance checks and testing proposals in REACH registration

dossiers8.

EFSA evaluations for food contact materials or food additives or other applications

(cosmetics etc.).

Literature data and data from public available databases including databases like:

- Public literature search information

- OECD toolbox9,

- CEFIC LRI Toolbox10 including RepDose, FeDTex and CEMAS

- Toxtree11 (TTC and related data)

- GESTIS substance database12,

- ChemIDplus (Toxnet, USA)13, Toxline14

- HPV-Program15,

- NICNAS (Australia)16

- Occupational Exposure Limits (OELs): see country specific lists

- Cosmetic Ingredient Review (CIR) database17), SCCS opinions

- Council of Europe Database 18 (not publicly available)

- PubChem19

- Chemspider20

6 Peer reviewed data sources are marked in bold

7 https://echa.europa.eu/addressing-chemicals-of-concern

8 http://echa.europa.eu/information-on-chemicals/dossier-evaluation-decisions

9 http://www.oecd.org/chemicalsafety/risk-assessment/theoecdqsartoolbox.htm

10 http://www.cefic-lri.org/lri-toolbox

11 https://eurl-ecvam.jrc.ec.europa.eu/laboratories-research/predictive_toxicology/qsar_tools/toxtree

12 http://www.dguv.de/dguv/ifa/Gefahrstoffdatenbanken/GESTIS-Stoffdatenbank/index-2.jsp

13 http://chem.sis.nlm.nih.gov/chemidplus/

14 http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen?TOXLINE

15 http://webnet.oecd.org/hpv/ui/SponsoredSubstances.aspx

16 http://www.nicnas.gov.au/chemical-information

17 http://www.cir-safety.org/ingredients

18 https://fcm.wiv-isp.be/

19 https://pubchem.ncbi.nlm.nih.gov/

20 http://www.chemspider.com

13

- ESIS21

- TSCA22

Read-across information from structural similar substances or chemical categories

The same search criteria as above mentioned apply (see further details under section 3.5)

Weight of evidence information (this involves assessing the relevance, reliability and

adequacy of each piece of available information, holding the various pieces of information up

against each other and reaching a conclusion on the hazard. This process always involves

expert judgement).

The same search criteria as above mentioned apply.

The reliability and relevance of the information collected has to be identified:

- GLP-studies vs. non-GLP- studies

- Klimisch rating (Klimisch 1997), further developed by Schneider et al.23

- ToxRTool24

- SCIRAP tool25

3.2 Uncertainty factors

Once it has been demonstrated that a substance does not pose any concern with regard to

genotoxicity, an appropriate dose descriptor from repeated dose (chronic/subchronic/sub-acute)

toxicological studies can be selected. Guidance on dose descriptor selection is for example available

through the ECHA guidance on information requirements and chemical safety assessment26. For

older dietary rodent studies, where only concentrations in feed are available, guidance is provided in

EFSA (2012)27.

A self-derived Tolerable Daily Intake (sTDI) can be derived from e.g. the NOAEL (No-Observed-

Adverse-Effect-Level), a benchmark dose (BMD-L) or – if these are not available - a LOAEL (Lowest

Observed Adverse Effect Level) (= the so-called Point of Departure - PoD) obtained from repeated

dose (chronic/sub chronic/sub-acute) toxicological studies and taking into account specific

uncertainty factors.

21

http://esis.jrc.ec.europa.eu/ 22

http://yosemite.epa.gov/oppts/epatscat8.nsf/reportsearch?openform 23

Schneider K, Schwartz M, Burkholder I, Kopp-Schneider A, Edler L, Kinsner-Ovaskainen A, Hartung T, Hoffmann S. (2009). “ToxR Tool”, a new tool to assess the reliability of toxicological data. Toxicology Letters 189(2):138-144 24

https://eurl-ecvam.jrc.ec.europa.eu/about-ecvam/archive-publications/toxrtool 25

Beronius A, Molander L, Ruden C, Hanberg A (2014). Facilitating the use of non-standard in vivo studies in health risk assessment of chemicals: A proposal to improve evaluation criteria and reporting. Journal of Applied Toxicology 34(6): 607-617 26

Chapter R8.2, the ECETOC report TR 85 - Recognition of, and Differentiation between, Adverse and Non-adverse Effects in Toxicology Studies; ECETOC report TR 99 - Toxicological Modes of Action: Relevance for Human Risk Assessment. 27

http://www.efsa.europa.eu/en/efsajournal/pub/2579

14

sTDI (mg/kg body weight/day) = PoD (mg/kg body weight/day) / uncertainty factor

According to recent guidance by EFSA (2012)28, the TDI is calculated by dividing the PoD obtained

from a repeated dose toxicity study with one or more uncertainty factors.

INTERSPECIES/INTRASPECIES UNCERTAINTY FACTOR

The interspecies/intraspeciel uncertainty factor has been set to be “100”. This factor gives an

additional margin to take into account the possibility that humans may be more sensitive than

animals and that some humans may be more sensitive than others.The factor 100 is constituted of

two factors of 10. One factor of 10 is intended to account for interspecies differences. This factor of

10 is envisaged as converting the findings in animals to equivalent findings in humans.

A second factor of 10 is used to account for differences in typical humans and sensitive sub

populations such as children, the elderly or compromised individuals.

It should be noted that slight differences exist between ECHA Guidance R8 and EFSA-imposed

interspecies extrapolation if the PoD is based on a test species other than the rat. This adaptation

accounts for differences in allometric scaling between the species. REACH DNELs might therefore

deviate from the estimated TDI.

These two uncertainty factors are intended to be conservative and address a wide range of

chemicals. Recent guidance provided by the International Program on Chemical Safety (IPCS) ,

ECHA and EFSA29 (2012) allows for deviation from the values of 10 when the data on the specific

substance is sufficient to justify alternative values. In certain instances, smaller values can be

justified using data on mechanism of actions or modeling of the pharmacokinetics of the compound.

If such a decision is made, this should be based on robust scientific justification.

STUDY DURATION UNCERTAINTY FACTOR

Extrapolation from short-term or sub-chronic-studies to lifetime exposure requires the application of

an additional uncertainty factor. According to EFSA (2012), a factor of 2 is sufficient to extrapolate

from a high-quality 90-day (sub-chronic) study to chronic exposure conditions.

28-day oral studies might be of use in the context of a weight-of-evidence approach. EFSA (2012)

suggests a case-by-case assessment to extrapolate to chronic exposure. It is noted that in case

substance-specific information is not available, ECHA suggests factor 6 as a default. This factor

accounts for the aforementioned factor 2 from 90-day to chronic exposure and an additional factor 3

for the extrapolation from 28-day to 90-day studies. This latter factor is based on the Rule of Haber

(effect = dose x exposure duration).

In some cases, developmental endpoints may provide the most sensitive PoD for risk assessment.

For example, these may be derived from a pre-natal development study (as shown below in the

example of 2-ethylhexanoate) or a 2-generation reproduction toxicity study. Since the exposure

28

http://www.efsa.europa.eu/en/efsajournal/pub/2579 29

http://www.efsa.europa.eu/en/efsajournal/pub/2579

15

duration covers the entire development of the offspring, developmental PoDs do not require the

application of an additional uncertainty factor for duration. In this context, it should be noted that the

NOAELs for parental toxicity in OECD 414, OECD 416 or OECD 443 studies usually do not provide

a preferable PoD, as the examination typically does not include important parameters used for the

assessment of systemic toxicity. Therefore, these should only be used if they provide the most

sensitive PoD.

NOAEL, LOAEL and BMD

While NOAELs and BMDs can be applied as such, the use of a LOAEL requires the use of an

additional safety uncertainty factor of up to 10. The magnitude of this factor should be based on the

overall dose-response and dose spacing in the experiments. Other regulatory guidances (ECHA R.8,

SCCS Notes of Guidance rev. 8 (2012) suggest a default factor of 3 unless data indicate otherwise.

ROUTE-TO-ROUTE EXTRAPOLATION

Oral studies provide the most relevant information for the safety assessment of food contact

materials. However, in some cases, oral studies are not available, or lack reliability or relevance for

risk assessment purposes. In such cases, studies with a different exposure route (typically

inhalation) may be used to provide the PoD for risk assessment as part of a weight-of-evidence

approach. Guidance on route-to-route extrapolation can be found in ECHA Guidance R.8.

With:

sRVrat = standard respiratory volume of the rat in a study with a daily exposure duration of 6 hrs =

0.29 m3/kg bw

ABS = Absorption (percentage of intake via a specific exposure route - needs either TK data or

modeling). Default assumption is 100 % in all listed cases.

Since route-to-route extrapolation underlies high uncertainties, any toxicokinetic information

available should be incorporated into the assessment. Highly refined approaches like biologically-

based toxicokinetic (PBTK) modelling may be applied if appropriate. Care should be given to effects

that may be route-specific (e.g. as indicated by information on similar substances).

UNCERTAINTY/DATA GAPS IN THE DATABASE

In those cases where available testing information is insufficient to provide information necessary for

a food contact safety assessment, additional uncertainty factors may be applied. This includes

hazard assessment by read across approaches if used as part of a weight-of-evidence approach.

Guidance on the application of uncertainty factors for read across approaches may, for example, be

found in Blackburn (2014).

16

TOTAL UNCERTAINTY

The total uncertainty is the product of all uncertainty factors used in the assessment:

Total UF = UFinterspecies/intraspecies x UFduration x UFLOAEL x UFdata gap/read across

For substances for which toxicity data are available, it is important to use all the data for selecting

the most appropriate NOAEL to determine the TDI. Only after all data, including uncertainty factors,

have been assessed, can the most sensitive and appropriate PoD for the TDI be selected.

Figure1 (next page) provides an overview of various options to derive a TDI for a substance and the

assessment factors to be applied. The Threshold of Toxicological Concern is discussed in the

following section.

17

.

18

3.3 The Use of the DNEL (REACH) for Risk Assessment in Food Contact

As explained under 3.1. and 3.2. the Tolerable Daily Intake (TDI) is derived as the safe dosage of a

substance migrating from food contact materials into food and which is therefore orally consumed via

the human diet. Thus, exposure from food contact materials only happens via the oral-gastro route.

In contrast, DNELs (Derived-No-Effect-Level) are derived for various routes (e.g. dermal, inhalation,

oral) and exposure groups (workers and general population). The REACH Regulation also describes

the methodology to define the PoD (point of departure, see 3.2.), the use of assessment factors and

the calculation of other DNELs based on DNELs in case of lacking of appropriate data.

The DNEL is a value described in the REACH Regulation under Article 119 Section 1f, and which

can also be used for finding a safe exposure limit of a substance in food contact materials. The

DNEL is defined under Annex I, Section 1.0.1 of the REACH Regulation:

“The objectives of the human health hazard assessment shall be to determine the classification of a

substance in accordance with Regulation (EC) No 1272/2008; and to derive levels of exposure to the

substance above which humans should not be exposed. This level of exposure is known as the

Derived No-Effect Level (DNEL).”

In accordance with the provisions laid down in the REACH Regulation, the registrant has to derive

such limits for all substances as of 10 tons/year (production or import) to gain quantitative values

based on the toxicological properties of that material. In general, the registrant does this in the form

of a self-assessment for all substances. The DNEL is described in detail in the ECHA –Guideline R8

Since DNELs are not officially approved or confirmed by ECHA (only checked according to a defined

procedure), an unreflective use for the risk assessment of NLS or NIAS from food contact materials

is not recommended. However, since there are clear rules in the ECHA-Guideline R8 for deriving

DNELs, these values are the outcome of an ECHA prescribed methodology and the current state of

the art. Nevertheless, it is advisable to realise on which data basis (oral, dermal or inhalative route)

and calculation method including safety factors the DNEL has been generated in order to be able to

take a decision whether this value is suitable or not suitable for food contact use. Therefore, it is

recommended to consult expert knowledge (toxicologist) to confirm the use of a DNEL. In some

cases, it might be useful to add additional uncertainty factors.

For the risk assessment of a substance the DNEL is used similar to the TDI and results in the

following two scenarios: If the ratio between level of exposure to DNEL is below 1, it is considered

safe (e.g. Exposure = 1 mg/kg Food and DNEL = 2 mg/kg Food = > ½ = 0,5). If the ratio between

level and DNEL is above 1, it is considered unsafe (e.g. Exposure = 10 mg/kg Food and DNEL = 1

mg/kg Food = > 10/1 =10). The ratio between DNEL (or TDI or uncorrected NOAEL etc.) and

exposure level is often defined as margin of exposure (MoE) or margin of safety (MoS). It is always

desirable to have a MoS or MoE of several magnitudes.

A straightforward Dose – Response – Correlation with a threshold mechanism is necessary to derive

a DNEL for a toxicological effect. The threshold marks the dose under which the toxicological effect

does not occur anymore. Effects which do not correlate with a specific dose cannot be used to derive

a DNEL. However, it is sometimes helpful to at least define a level where the occurrence of an

adverse effect is minimal. The so-called DMEL (Derived-Minimal-Effect-level) corresponds to a

Dose-Response-relationship without a threshold. The DMEL should only be used as an exception

19

and only together with appropriate uncertainty factors (e.g. for risk assessment of a carcinogenic

residue present in the product in trace levels).

In general, DNELs for REACH-registered substances are publically available on the ECHA-website:

Registered substances .For food contact uses the preferred DNEL to be used is “General Population

- Hazard via oral route – systemic effects – long term exposure” (hereinafter referred to as

DNELGP_oral_longterm). If this DNELGP_oral_longterm is calculated based on solid tox data (e.g. PoD, NOAEL

or LOAEL) and appropriate uncertainty factors (the terminology “assessment factor” is used the

REACH context), the value can directly be used for the Risk Assessment of NLS or NIAS. The

DNELGP_oral_longterm should be based on chronic feeding studies with the original substance; if this is

not the case always seek for toxicological support to determine how valid the figure is. Cross

calculation using read across data or other DNEL (e.g. based on inhalation or dermal) may be not

valid enough to be used for food contact application.

A value for the used “overall assessment factor” can also be found in the REACH Dossier for the

substance. This factor is a combined value used to calculate the DNEL from a point of departure,

such as a NOAEL (No-observed-adverse-effect-level), NAEL (No-adverse-effect-level) or a LOAEL

(lowest-observed-adverse-effect-level). The uncertainty factors are adapted according to the studies

available. The REACH guideline for DNELs advises which assessment factors should be combined.

The NOAEL levels etc. are in general the outcome of animal studies and the various factors (incl.

safety factors) are used to translate the value into a safe dosage for humans. If the approach of

deriving a DNELGP_oral_longterm using many uncertainty factors appear too conservative it may be easier

to refer to the TTC-Concept and use these generic exposure limits based on the criteria applied

there (e.g. structure, no-CMR, etc.).

3.4 Read across

READ ACROSS (Q)SAR

When testing data are insufficient or lacking, toxicological properties can be read across from toxicity

information of similar substances. If the prediction is based on modelling, the term (Q)SAR

(quantitative structure-activity relationship) would apply, with a toxicological activity being the

function of one or more physicochemical parameters of the substance. Detailed guidance for toxicity

prediction is beyond the scope of this guidance; however, some key elements are summarised to

improve the understanding. For any prediction of a toxic effect, with or without computational

support, the following principles should be considered:

Sound expertise in both toxicology and chemistry is required for a critical appreciation of the

results.

Any prediction requires complete and transparent documentation. Ideally, the level of

documentation should be equal to that provided in a QMRF (QSAR Model Reporting Format

and QPRF (QSAR Prediction Reporting Format) (ECHA, 2012)

In general, results of actual testing are considered to be of higher relevance for risk

assessment; only in case of severe doubts of the reliability of the existing test may non-

testing data be applied as part of a weight of evidence approach.

The (Q)SAR model applied should fulfil the validity criteria outlined by OECD (2007), which

are defined as follows:

20

- A defined endpoint: A defined endpoint might be mechanistic (e.g. estrogenic activity), a test

protocol endpoint (e.g. liver weight) or regulatory (NOAEL). Clear distinction should be made

on the applicability of the endpoint to risk assessment).

- An unambiguous algorithm: Thorough documentation should ensure that the rationale

underlying the prediction is transparent and the results are reproducible.

- A defined domain of applicability: For example, a model based on hydrocarbon data should

not be used for organic acids. The chemical space is defined by a border of physicochemical

parameters that should always be checked before a model is applied.

- Appropriate measures of goodness-of-fit, robustness and predictivity: Models should be

statistically robust, and the predictor variables should be chosen to be independent from

each other (e.g. not using both boiling point and vapour pressure, or molecular weight and

size).

- A mechanistic interpretation, if possible: The model should ideally be plausible, with a causal

link between physicochemical property of the chemical substance and biological effects

existing. For statistical models, this might not always be possible.

For the assessment of food contact materials, genotoxicity/carcinogenicity and repeated dose

systemic toxicity are the key hazard classes of interest and are described in more detail.

a) Genotoxicity / Carcinogenicity

The TTC approach requires a decision of presence or absence of the potential for (direct)

genotoxicity. The reactivity with DNA is dependent on the presence of electrophilic functional

groups (structural alerts), and various computational tools exist to predict whether a molecule

is likely to bind to DNA. Computational tools can either be knowledge-based (rule-based)

(providing information on the existence of known structural alerts), statistical (comparison

with test results of similar molecules, e.g. with identical fragments), or hybrids of these two.

The use of in silico methods to confirm the absence of structural alerts for genotoxicity is

accepted in several regulatory frameworks (EFSA, 2012; ICH, 2014), but regulators do not

provide clear guidance on how to confirm the absence of a structural alert. However, some

considerations may offer support:

The Benigni-Bossa rule base in Toxtree is reported to predict genotoxicity with an

accuracy of 80 % (Benigni, 2008), which is comparable to the Ames test.

Various publications suggest several models to be used in parallel, to reduce false

negative predictions. The increased probability of false positive predictions caused by

that approach is accepted based on the fact that the TTC is meant to provide a

“safety net”, and a false positive would only result in an unnecessary Ames test to be

conducted.

For impurities in pharmaceutical actives, regulators provided a more defined

guidance: “The absence of structural alerts from two complementary (Q)SAR

methodologies (expert rule-based and statistical) is sufficient to conclude that the

impurity is of no mutagenic concern, and no further testing is recommended” (ICH,

2014)

21

The OECD QSAR Toolbox is a meta-database, including key elements of Toxtree,

OncoLogic and additional lists of strong electrophilic groups with potential DNA

reactivity. It also includes both a rule-based and a statistical element.

Note that some substances (e.g. chloroform, CAS# 67-66-3) may induce carcinogenicity in rodent

bioassays in the absence of any genotoxic mode of action. Typically, non-genotoxic carcinogenicity

is a secondary effect of target organ toxicity or endocrine activity and underlies a dose-response

relationship, including a NOAEL that can be used for risk assessment.

b) Repeated dose systemic toxicity

Due to great uncertainty in both toxicokinetics and toxicodynamics, non-genotoxic systemic

oxicity has shown to be the biggest challenge in predictive toxicology. Computational (Q)SAR

models to predict systemic toxicity underlie rapid development and are not further described

here. Examples may be found in Schilter (2014).

Regulatory bodies differentiate between read across from one of few analogues or from a

group/category of chemicals. Detailed guidance has been published by regulators (ECHA,

2008; ECHA, 2015 and WHO, 2014), joint initiatives such as ILSI (Schilter, 2014), SEURAT-1

(Schultz, 2015) and by industry (see e.g. Wu, 2010; Blackburn, 2011; Blackburn, 2014,

Patlewicz, 2015).

The figure below shows three possible ways how prediction of toxic effects can be applied:

Property 1/Activity 1 (read across/SAR): Both properties (log Kow, electronegativity,..) and

activities (e.g. covalent binding to DNA or proteins, ligand receptor interaction) may be read

across from one source substance to a target substance (Chemical 1->2, Chemical 3 -> 4)

Property 2/Activity 2 (Interpolation): Properties of the target substances (Chemicals 2 and

3) are predicted from properties within the chemical space of a group of source substances

(Chemicals 1 and 4). This means, the group consists of source substances that are expected

to be either less (e.g. Chemical 1) or more (e.g. Chemical 4) lipophilic, electronegative, DNA-

reactive, hydrolytically unstable than the target substances. All predictions of properties or

activities stem from interpolation within a defined group.

Property 3/Activity 3 (Extrapolation): In this example, also a group (category) approach is

applied. Source substances (Chemicals 2 and 3) are not believed to represent the

boundaries for specific properties or activities, but the predictions for the target substances

(Chemicals 1 and 4) are believed to be outside the chemical space of the group.

22

Figure 2. Graphical representation of a chemical category and some approaches for filling

data gaps (OECD, 2014)

In general, the category (group) approach is regarded to be more robust than read across from

one analogue substance, and interpolation (within a defined applicability domain/chemical

space) is believed to be more reliable than extrapolation to a target substance outside the

boundaries defined by the properties of the source substance(s).

Typically, the rationale for read across is a common mode of action, identical metabolites,

and/or strong similarity of the parent molecules (both based on physicochemical properties and

presence/absence of specific functional groups). The OECD QSAR Toolbox may aid in the

search for a suitable category for read across.

Other strategies for the definition of a category or the search for an analogue may include a

similarity search via the Tanimoto Index (e.g. via www.chemspider.com, Toxmatch or the OECD

QSAR Toolbox). This should always include a prediction of potential metabolites

Toxtree and the QSAR Toolbox are freely available examples for metabolism prediction.

Current research focuses on the identification of Adverse Outcome Pathways (AOP) that will aid

to break down the complexity of systemic toxicity to a sequence of simple steps that are easy to

model.

While a detailed guidance is beyond the scope of this document, some examples will highlight

the uncertainty/possible pitfalls related to read across:

Hydrocarbons/hexane: Short-chain aliphatic linear hydrocarbons (e.g. C2 to C8) may form a

homogenous group of chemicals with similar toxicological properties. However, hexane has

been found to be an outlier within this group, inducing chronic and reproductive toxicity that is

unique to this compound. Therefore, even interpolation within a group of very similar category

members should be interpreted with care.

23

Benzene/Toluene/Ethylbenzene/Xylenes: Even though these molecules differ by only one

methyl group, each of the mentioned substances induces unique target organ effects that have

not been observed with any of the other compounds. While these molecules seem to be very

similar to each other, prediction of their metabolic breakdown would show the formation of

different degradation products, resulting in great variation of toxicity.

Benzophenone/4-methylbenzophenone: In its risk assessment of the finding of 4-

methylbenzophenone in breakfast cereals in 2009, EFSA found that there was very little

information useful for the toxicological assessment of 4-methylbenzophenone itself, but there

was much more information available for the structurally similar benzophenone. 4-

Methylbenzophenone is expected to be metabolised by the same metabolic pathways as

benzophenone, with the addition of oxidation of the 4-methyl group to the corresponding

alcohol and further oxidation to the carboxylic acid with its glycine and glucuronide conjugates.

Based on structural considerations and experimental results on the structurally related

benzophenone, it can be concluded that 4-methylbenzophenone does not raise concern for

genotoxicity, and like benzophenone, 4-methylbenzophenone is expected to be a non-

genotoxic carcinogen. In addition to the usual uncertainty factor of 100 to allow for inter- and

intraspecies differences in sensitivity, EFSA applied a further uncertainty factor of 2 to allow for

the read across of data from benzophenone to 4-methylbenzophenone.

In conclusion, read across of systemic toxicity is related to great uncertainty and should be done with

care. However, this field is subject to broad interest and rapid development.

3.5 Determination Threshold of Toxicological Concern (TTC)

In cases where no or insufficient animal testing data are available for a chemical substance and no

read-across is possible, pragmatic approaches to determine acceptable threshold limits have been

developed over more than four decades. In particular, the threshold of toxicological concern (TTC) is

a concept defining exposure thresholds for substances below which no appreciable risk for human

health is assumed (Kroes et al. 2000, 2005; Kroes and Kozianowski 2002).

The TTC concept has been used for years in the assessment of impurities of food contact materials,

drinking water, and pharmaceuticals (for a comprehensive overview of existing values for the TTC,

see Hennes 2012). For example, the US Food and Drug Administration (FDA) adopted the concept

of a threshold of regulation (TOR) for substances used in food contact articles: If a substance or an

impurity has not been shown to be a carcinogen in humans or animals and there is no reason, based

on the chemical structure of the substance, to suspect that it is a carcinogen, a threshold of

regulation is defined as a dietary concentration of 0.5 ppb (=μg/kg diet) or 1.5 μg/person/day

assuming a consumption of 3 kg diet per day. The threshold was derived from linear extrapolation of

TD50 values from animal experiments to the risk of one in a million for tumor development in humans.

Other concepts have further developed the general TTC approach, based either on the structure of

groups of substances (Munro et al. 1996; Kroes et al. 2004) or on information on certain endpoints

(Cheeseman et al. 1999) which would allow setting higher threshold concentrations under certain

conditions with a high degree of certainty to predict that a substance is safe if the concentration is

not exceeded. Based on an analysis of a comprehensive database of 2,944 entries for 600

substances on repeated dose studies and studies on reproductive toxicity and developmental

24

toxicity, Munro et al. (1996) proposed human exposure thresholds for three structural classes as

defined by Cramer et al. (1978) using the 5th percentiles of the NO(A)ELs based on the lowest

NO(A)EL for each substance.

The human exposure thresholds were 1,800, 540, and 90 μg/person/day for Cramer class I, II, and

III respectively.

An overview of the most important threshold values (Figure 3):

Classification TTC

(µg/person/d)

TTC(µg/kg

b.w./d)

Reference

Cramer class I 1800 30 Munro 1996

Cramer class II 540 9 Munro 1996

Cramer class III 90 1.5 Munro 1996

Organosphosphates and carbamates 18 0.3 Kroes 2004

No structural alerts for

carcinogenicity

1.5 0.025 TOR

Genotoxic substances (without

aflatoxin-like, azoxy- or N-nitroso

compounds)

0.15 0.0025 Kroes 2004

In 2012 the European Food Safety Autority (EFSA) published a scientific opinion on exploring

options for providing advice about possible human health risks based on the concept of Threshold of

Toxicological Concern (TTC):

The Scientific Committee concluded that the TTC approach should not be used for the following

categories of substances: high potency carcinogens (i.e. aflatoxin-like, azoxy- or N-nitroso-

compounds, benzidines, hydrazines), inorganic substances, mixtures of substances containing

unknown chemical structures, metals and organometallics, proteins, steroids, nanomaterials,

radioactive substances, and substances that are known or predicted to bioaccumulate. Furthermore,

special attention has to be paid if there is data showing that a substance has endocrine-mediated

adverse effects and if the TTC approach is used for risk assessments including infants and children

the low body weight has to be taken into account.

The TTC value of 0.15 μg/person per day, derived by Kroes et al. (2004) for substances with a

structural alert for genotoxicity, is considered to be sufficiently conservative to be used in EFSA’s

work, excluding high potency carcinogens (see above). Possible genotoxic metabolite has to be

considered. Non-genotoxic carcinogens are considered to have a threshold and, in general, NOELs

for these are in the same range or higher than NOELs for other types of toxicity.

Also the TTC value of 18 μg/person per day, first proposed by Kroes et al. (2004), is considered

sufficiently conservative to cover the anti-cholinesterase activity of substances with

organophosphate or carbamate structural features. The Scientific Committee concluded that the

original FDA Threshold of Regulation (TOR) value of 1.5 μg/person per day is of historical

importance, but has little practical application in the overall TTC approach. However, it should be

25

noticed that this threshold value is the only threshold value really established by law (21 CFR (FDA),

§170.30) until now.

Only a very low number of substances formed the basis for the calculation of the Cramer class II

threshold value. Therefore, the Scientific Committee concluded that consideration should be given to

treating substances that would be classified in Cramer Class II under the Cramer decision tree as if

they were Cramer Class III substances.

All in all, the Scientific Committee concluded that the science supports the application of the TTC

approach in any area of chemical risk assessment for which human exposures are low, whether

exposure is from deliberate addition or due to contamination. Within EFSA, the Scientific Committee

recommends that the TTC approach can be used to assess impurities, breakdown and reaction

products, metabolites, and low-level contaminants in food and feed, where an exposure assessment

can be conducted, but on which there are few or no toxicological data. Therefore, the TTC approach

can be recommended as a useful screening tool either for priority setting or for deciding whether

exposure to a substance is so low that the probability of adverse health effects is low and that no

further data are necessary (EFSA Journal 2012;10(7):2750).

EFSA proposed the following generic scheme (Figure 4):

26

*Exclusion categories: high potency carcinogens, inorganic substances, metals and organometals, proteins,

steroids, substances known/predicted to bioaccumulate, nanomaterials, radioactive substances, mixtures

**If exposure of infants < 6 months is in range of TTC -> consider if TTC is applicable

***If exposure only short duration -> consider margin between human exposure and TTC value

Please notice, that a proposed revised decision tree was published on the 10th of March 2016 in the EFSA

report “Review of the Threshold of Toxicological Concern (TTC) approach and development of new decision

tree.”

EFSA recently reviewed the TTC approach and proposed a new version of the TTC decision tree

(EFSA event report, 2016). In this new decision tree the threshold levels are the same as given

above, besides the reintroduction of the Cramer class II value of 9 µg/kg bw/day. Despite

acknowledgement that there are very few chemicals in Class II and therefore the TTC value for this

class is not well supported within the current TTC approach, and the previous proposal to evaluate

under the Class III TTC threshold all the chemicals categorized as Class II (EFSA, 2012), the expert

group recommended that Cramer Class II continues to be used and applied to the TTC approach. A

change of wording and a change in the order of the asked questions were suggested. Furthermore, it

was stated that a merge of different non-cancer databases is desirable to increase statistical power

and improve transparency in the database, and that after the merge of the databases the “overall

TTC’s” should be recalculated. Thus, the TTC approach is still under revision and further changes of

the TTC decision tree should to be expected in the future.

27

Currently the TTC approach is used by several authorities like EFSA and FDA for risk assessment of

substances. However, there are several limitations for the use of the approach, which have to be

considered. The chemical structure of the compound has to be known, specific substance classes

are excluded (see above) and the TTC approach will only be used if there is no useful animal testing

data and no read across is possible.

Despite these limitations, there are considerations made about whether or not the TTC concept

could also be used for the evaluation of NIAS. Because the chemical structures of the NIAS are

often not known there is a difficult discussion which is still ongoing. In 2011, Koster et al. proposed a

TTC approach for the regulation of unknown substances found in food samples including non-

intentionally added substances (NIAS), but the confident identification of NIAS and, particularly,

genotoxic substances remains an unresolved issue. A step wise approach was suggested for the

evaluation of NIAS:

Several published studies have demonstrated that the application of the Cramer classification

scheme in the TTC approach is in general conservative and therefore protective of human health.

However, it should be pointed out that substance specific toxicological data or the read across to

28

similar substances would always be preferred, and that the use of the approach demands a high

level of Expert Judgement. All in all, the TTC concept is of great utility and of high interest due to the

pragmatic approach, conservativeness, and applicability in several areas. Furthermore, the use of

the approach for risk assessment of specific substances or as a screening tool for the prioritisation of

toxicological testing contributes to save animal lives and high costs.

3.6 10ppb- approach / 50 ppb- approach

The “10 ppb threshold” utilised in Europe is a limit of detection for the validated analytical

determination of migrants in food or food stimulant (0.01 mg/kg). Under the EU Regulation 10/2011

for plastics it is referred to in article 13 (specific provisions for plastic multi-layer materials and

articles): a substance whose migration is not detectable with a detection limit of 0,01 mg/kg is

exempted from authorisation, when it is in a layer not in direct contact with food, it is not classified as

‘mutagenic’, ‘carcinogenic’ or ‘toxic to reproduction’ category 1A, 1B or 2 according to the CLP

Regulation 1272/2008 and it is not in nano form. In addition, the detection limit of 0.01 mg/kg applies

to substances listed in annex I and which are subject to a non detectable migration limit (including

among others also carcinogenic substances like ethylene oxide and propylene oxide). However, it

should be pointed out, that in such cases, when the analytical methods enable a determination even

below 10 ppb, this has to be considered to be the applicable detection limit.

Furthermore, the 10 ppb concept has been introduced in industry code of practices and in national

legislation dealing with non plastics materials. Examples are the Dutch Commodities Act Regulation

on packagings and consumer articles coming into contact with foodstuffs (Warenwet) which allows

the use of non-listed components provided they are not CMR classified and their migration is less

than 0,01 mg/kg, or the CEPE code of practice (edition 4, version 9) for coated articles where the

food contact layer is a coating and which allows the use of monomers and additives provided they

are not CMR classified and their migration is less than 0.01 mg/kg . The Swiss Ordinance on

packaging ink on non food contact side (SR 817.023.21, section 8b) also allows the use of

substances which are listed but not yet toxicologically evaluated (part B of the lists), provided their

migration is not detectable with a detection limit of 0.01 mg/kg.

Lastly the detection limit of 0.01 mg/kg food is widely used as a cut off limit by testing laboratories for

the assessment of food contact materials, especially during screening tests (e.g. by GC-MS) on

extracts/migration solutions. The recently published ILSI guidance on best practice on the risk

assessment of NIAS in FCMs recommends the 10 ppb threshold during the non targeted chemical

analysis step (section 5.1). Its use is however conditioned to the exclusion of CMR substances

based on expert judgment or otherwise.

In the US a similar approach is used by industry and law firms, working as consultants and

performing substance evaluations for food contact applications. Due to this approach an analytical

sensitivity level of 50 ppb, a finding of “non-detected” is found to be reasonable. This limit value is

based on the “Ramsey Proposal”, a notice of proposed rulemaking circulated by FDA in 1969, and a

precursor and historical part of the development of the Threshold of Regulation (TOR) and the TTC

concept. Although never adopted by FDA, this concept has received wide acceptance in the

scientific community. Furthermore, due to the definition of a food additive as: “any substance the

intended use of which results or may reasonably be expected to result, directly or indirectly, in its

becoming a component or otherwise affecting the characteristics of any food (including any

substance intended for use in . . . . . . packaging, . . . or holding food . . .).”, it can be assumed that a

29

food-contact substance must be expected to become a component of food in more than a

toxicologically insignificant amount to be properly considered a food additive. An analytical sensitivity

level, a “non-detection” level of ≤ 50 ppb can be seen as a toxicologically insignificant amount for

substances for which a genotoxic potential does not have to be expected due to missing genotoxic

structural alerts or in vitro genotoxicity studies demonstrates that the substance is not genotoxic.

FDA's tiered requirements for toxicological data for substances with a potential dietary exposure

between 0.5 ppb and 50 ppb mandate two in vitro genotoxicity studies demonstrating that the

substance is not genotoxic. Also in Europe when submitting a petition for the listing of substances to

EFSA for food contact applications the requirements regarding the toxicological data package are

depending on the specific migrating of the substance into food simulants in the used application. If

the specific migration is found to be below 50 ppb only a reduced data set of three in vitro

genotoxicity tests has to be delivered. These set limits also support the use of a “50-ppb approach”

as a limit below which no risk of harm has to be anticipated if a lack of genotoxicity has been shown

by in vitro testing or no genotoxicity testing is necessary due to the lack of structural alerts based on

expert judgement.

In conclusion, if it can be shown that the migration of a substance, which is not CMR classified and

shows no genotoxic structural alerts, is less than 0.01 / 0.05 mg/kg food (simulant), then no further

assessment is needed. These conservative approaches have however the drawback of a very

limited applicability, due to the low values (EU: 10 ppb, US: 50 ppb). As a next step following a tiered

approach the possibilities of applying the TTC concept can be considered (see chapter 3.8).

Suggested step wise approach for the evaluation of non-listed substances/impurities present at low

concentrations/amounts (Figure 5):

30

3.7 Nano

According to the EU definition, engineered nanomaterial means a manufactured or processed

material in an unbound state or as an aggregate or an agglomerate where one or more external

dimension is in the size range 1 nm – 100 nm. These particles can be present in food contact

materials as IAS or NIAS.

According to EFSA, the safety of nanomaterials in FCM should be evaluated ‘case-by-case’ as there

is uncertainty of characterisation, detection and measurement of the nanoparticles in food and these

materials have specific properties, which may affect their toxicokinetic ant toxicology profiles. If it can

be convincingly demonstrated, that there is no migration of the nanomaterial to the food matrix then

there is no exposure to the substance and thus no toxicological concern. EFSA has used this “no

31

exposure – no risk”-concept in the approval of TiN nanoparticles (FCM #807) as food contact

material.

3.8 Endocrine Disruptors

While as of today no legal definition is in place, the European Commission presented in June 2016

criteria to identify endocrine disruptors in the field of plant protection products and biocides. The

Commission proposed to adopt a science-based approach to the identification of endocrine

disruptors and to endorse the WHO definition. In 2013, EFSA published a Scientific Opinion on the

hazard assessment of endocrine disruptors (EFSA, 2013). EFSA supported the OECD conceptual

framework substances testing of potential endorcrine disrupting properties, which is based on the

presence of an adverse effect caused by an endocrine mode of action. EFSA concluded "EDs can

therefore be treated like most other substances of concern for human health and the environment,

i.e. be subject to risk assessment and not only to hazard assessment." However, hazard

assessment should pay attention to possible gaps in the test data set. ECETOC (2009) suggests the

application of additional uncertainty factors in some cases.

3.9 Bioaccumulation / bioavailability in the human body

Another major aspect with respect to risk assessment is the molecular weight of a substance or in

case of polymers the fraction < 1000 Dalton. Substances with a molecular weight > 1000 Dalton are

unlikely to pass biological membranes meaning they are non-bioavailable and therefore they are not

expected to cause adverse systemic effects, unless they hydrolyze in the gastrointestinal tract and

liberate toxic substances or cause a local effect like irritation of the mucosa. If this can be excluded,

only substances or oligomers with a molecular weight < 1000 Dalton normally need a risk

assessment.

Bioaccumulation in the human body occurs when a substance is absorbed by the body to a higher

extent than it is excreted. As substances with a molecular weight > 1000 Dalton are normally not

absorbed by the body, bioaccumulation cannot occur. Thus, this toxicological endpoint is relevant

only for substances < 1000 Dalton.

Extract from the note for Guidance, annex 4 to Chapter III: In the case of food contact materials the

interest centers on the potential for direct accumulation in mammalian tissues and not on

biomagnification through the food chain. However, normally a log ko/w value below 3 would be

considered sufficient evidence for the lack of accumulative potential in the mammalian body, unless

special considerations, e.g. chemical structure, give cause for concern. On the other hand, a log

ko/w of 3 and higher will not by itself be proof of accumulation as a substance may not be absorbed

or be metabolised to substances with no accumulation potential. In these circumstances other

evidence for the absence of accumulative potential is needed.

32

4. EXPOSURE ASSESSMENT

Exposure assessment aims to define the dose of non-listed substances that individuals receive in

exposed populations. This dose is the so-called Estimated Daily Intake (EDI) (mg/person/day).

The EDIs for non-listed substances in food contact materials are estimated in a number of ways

depending on the material and the nature of the contact. To assess the dietary exposure to a

substance migrating from “repeated-use” applications (e.g. pipes, tubing, food containers, and food

processing equipment) conservative models are applied. The assessment of food contact materials

that come in contact with foods in non-repeated use applications e.g. food packaging is more

complex and often requires more refined models and additional data. In both cases, tiered

approaches are typically used in exposure assessments. Tiered approaches begin by using simple,

conservative, and widely applicable models of exposure. These models require relatively little data

but tend to overestimate exposures. If the exposure estimates are found to be too large using the

conservative models then the assessor moves on to more refined methods.

All exposure assessments for non-listed substances require the same types of information. These

include data on the ability of the substance to migrate from the material into food or water during

contact events (migration data), and data that allows the prediction of the daily dose to exposed

individuals (food consumption and food packing data). The findings of migration are a property of the

substance, the food contact material, the food, the duration and conditions of the contact

(Temperature, S/V). The findings of exposure are determined by how much food and water are

consumed by the consumer and what types and shapes of packaging are used for the food and

water.

4.1 Tiered approach

The consumer exposure to a given migrant can be determined by various basic to complex means,

depending on how much refined value is needed for the risk assessment, i.e. depending on

substance hazard. Consequently, a tiered approach is proposed, starting with a simple worst case

determination and ending with a refined determination which requires analytical data and knowledge

of intended uses (FCM structure, storage conditions, food categories and consumption data…). The

latter approach is more difficult to conduct at an early stage in the supply chain e.g. for a food

contact additive manufacturer, and the exposure data is then only relevant to very specific uses of

the product.

In any case, the exposure assessment necessarily first requires the determination of the migration

level of substance of interest from intended FCMs (refer to paragraph 4 b), either by calculation (

100% transfer or migration modeling) or by a combination of extraction data (residual content in

intermediate or final material) and calculation or by migration tests on final material.

Migration values have then to be converted into consumer exposure (amount of substance

/pers/day) for comparison with “tolerable intake values”. Here also a tiered approach is proposed to

estimate the “FCM consumption” (expressed e.g. as dm2 /pers/day), refer to paragraph 4d.

33

4.2 Migration Estimation

The amount of migration of substances from Food Contact materials into food can be derived from

worse case calculations (assuming 100% migration), migration models (diffusion model), extraction

studies in solvent (experimental data) or migration studies in food simulants (experimental data).

Migration is normally expressed in mg/dm2 Food Contact material or in mg/kg food.

WORSE CASE MIGRATION

The following formulas can be used to calculate the worse case migration in the food assuming

100% migration:

Option 1

CPolymer (mg/kg) * dPolymer (g/cm3) * SPackaging (cm2) * ePackaging(cm)

CFood (mg/kg) = ------------------------------------------------------------------------------------

M Food(g)

Where:

CPolymer: concentration of the substance in the polymer

CFood: concentration of the substance into the food

dPolymer: density of the polymer

ePackaging: thickness of the packaging material

SPackaging: contact area of the packaging material

MFood: weight of the food in contact with the material

Option 2

CFood = Cpolymer x [dPolymer / dFood ] x [SPackaging / VFood] x ePackaging

CFood (ppm) = Cpolymer (ppm) x [dPolymer / dFood] x 0.1 x [SPackaging / VFood](dm-1) x 104 x ePackaging (µm)

Where:

CFood = concentration of the substance into food (mg/kg or ppm)

CPolymer = concentration of the substance into polymer (mg/kg or ppm)

dPolymer = density of the polymer (g/cm3 or kg/dm3)

SPackaging = contact area of the food contact material (cm²)

ePackagingM = thickness of the food contact material (cm or 104 µm)

VFood = volume of food (cm3)

dFood = density of the food (g/cm3)

* NOTE: Difference between options 1 and 2 are units

34

EXAMPLES

Printing Inks. In case the substance is used in printing ink applications, EuPIA (European Printing in

Association) has published some default values which can be used to calculate the worse case

migration30.

When substances are used in printing inks an average ink weight per area is required to be able to

calculate the worse case migration. The following dry ink film weight (indicative values) can be used.

Printing technology Ink film weight

(dry)

Flexographic ink 1-1.5 g/m²

Gravure ink 1-2 g/m²

Offset ink 1-2 g/m²

Dispersion varnish 2-3 g/m²

White basecoat 12-16 g/m²

Clear basecoat 1-2 g/m²

UV varnish 4-7 g/m²

The following formula can be used to calculate the worse case migration in the food assuming 100%

migration originating from the printing ink.

W (g/m2)* CInk (%)* SPackaging (dm2)

CFood (mg/kg) = ------------------------------------------------------

M Food(kg) * 10

Were:

CFood: concentration of the substance into the food

CInk: concentration of the substance in the dried ink

MFood: weight of the food in contact with the material

SPackaging: contact area of the packaging material

W (g/m2): Dry Ink weight of the printed article

Coatings. CEPE (European Council of Paint, Printing Ink and Artists’ Colours Industry) describes in

its “Code of practice for coated articles where the food contact layer is a coating” the typical film

thickness used in different coating applications31

Application Film Thickness

Coated light metal

packaging

5 – 15 µm

Drums and Pails 12 – 13 µm

Heavy duty coatings 250 – 500 µm

30

http://www.eupia.org/uploads/tx_edm/2011-11-14_EuPIA_Guideline_for_Food_Packaging_Inks_-_November_2011__corr_July_2012.pdf 31

http://www.cepe.org/epub/easnet.dll/ExecReq/Page?eas:template_im=100087&eas:dat_im=05043D

35

Adhesives. In May 2016, the Association of the European Adhesive & Sealant Industry (FEICA)

published a guidance paper called “Migration testing of adhesives intended for food contact

materials”32. The guidance paper forms part of a package on migration testing of non-plastic food

contact materials developed by several sector associations from the packaging supply chain.

Due to the wide range of applications and the complexity of the chemistry, no unified testing

conditions can be defined for adhesives. Testing adhesives according to the rules of the plastics

regulation without substrate or the construction material will usually overestimate the migration of

constituents into foodstuff, as contributing factors to real migration are not sufficiently considered.

Contributing factors can be:

Curing times and conditions

Interaction of adhesive with other FCM layers

Barrier properties of other FCM layers

Distribution of constituents within the FCM

Ratio of adhesive amount to filling good

The Guidance includes material-specific properties to be considered when testing. Also, general

migration testing recommentations for i.e typical application in paper and board packaging, as well

as specific recommendations for pressure-sensitive adhesives and cold and heat seal application

are included.

EXTRACTION TESTS

Another way of obtaining information on the possible worst case migration of substances from food

contact materials is to perform extraction tests on the food contact material with an appropriate

solvent. By applying extraction test information on the possible migration can be obtained relatively

easier than by performing migration studies. However it has to be ensured that a suitable extraction

solvent is used obtaining a 100% extraction of the component of interest.

It has to be pointed out that from the results obtained from extraction studies non-compliance may

not be concluded. Actual migration studies or analysis in packed food itself are required.

MODELING

For predicting the migration of substances, mathematical modeling can be applied, which has been

significantlydeveloped in recent years. These tools have been validated for some of the common

used plastics and provide an over estimation of the possible actual migration. For guidance on

migration modelling JRC (Joint Research Centre) issued a guidance document33.

32

http://www.feica.eu/information-center/news/feica-guidance-paper---migration-testing-of-adhesives-intended-for-food-contact-materials.aspx 33

http://publications.jrc.ec.europa.eu/repository/bitstream/JRC59476/reqno_jrc59476_mathmod_v10_cs_2010_09_24_final.pdf%5b1%5d.pdf

36

Modeling on plastics has been accepted by EFSA as an option to calculate migration34. Modeling is

only applicable under “non-swelling” conditions. For other materials, like paper and paperboard the

development of a modeling tool is in progress.

Modeling is not yet applicable for rubbers or other elastomers.

It is known that in most cases the migration of substances from polymeric materials follows Fick’s

law of diffusion. In order to be able to use the modeling tools at least the following information is

needed; molecular weight of the substance with a polymer specific parameter (Ap) and a polymer

specific activation energy (t) to obtain the diffusion coefficient D of the organic substance in the

polymer, amount of substance in the polymer, layer thickness, area to volume ratio and the contact

condition of the food contact material.

SOME AVAILABLE SOFTWARE TOOLS (non-exhaustive list of tools)

There are a few companies who offer software systems for migration modelling such as: INRA Safe

Food Packaging Portal version 335, MIGRATEST software36 or AKTS-SML Software37, FACET

v3.0.0, among others.

NOTE: The first three are designed to overestimate migration and are valiadated, while FACET

offers a more realisitic calculation but is not validated.

MIGRATION TESTS

The European Commission together with the Joint Research Center (JRC) is currently finalising the

“Technical Guidelines for migration testing compliance”. These technical guidelines are part of a

series of documents to provide guidance on application of Regulation (EU) No 10/2011 on plastic

materials and articles intended to come into contact with food and are only applicable to plastic

materials and articles in the scope of this Regulation. It covers the following topics related to

compliance testing of plastic materials and articles: sampling, testing in food, choice of food

simulants and test conditions, testing in food simulant, verification testing, screening testing,

calculation of migration test results and reporting of test results. Although these technical guidelines

are not legally binding, they give the state-of-the-art of compliance testing in the framework of

Regulation (EU) No 10/2011 and they give all the elements for the supporting documentation related

to the compliance testing of the material or article. Once finalised, the guidelines will be made

available on the EC’s website.

In parallel industry stakeholders associations, covering the so-called “non-plastics”- value chain have

formed a task force to propose testing conditions better adapted to the specificity of various FCMs.

The Task Force is developing compliance guidelines with separate chapters for each non-

harmonized FCM. The guidelines explain that in the Plastics Regulation (EU/10/2011) some of the

simulants, times and temperatures are inappropriate for some non- plastic FCMs. However in the

absence of harmonised regulations, the conditions used in the Plastics Regulation are often applied

to non-plastics. Plastic simulants and/ or conditions and may cause physical damage or changes to

34

http://www.efsa.europa.eu/sites/default/files/scientific_output/files/main_documents/3635.pdf 35

http://modmol.agroparistech.fr/SFPP3/SFPP3download.html 36

http://www.fabes-online.de/software.php?lang=en&mode=migratest 37

http://www.akts.com

37

the non-plastic FCM leading to wrong results. Each sector assesses the applicability or not of the

plastics 10/2011 migration testing guidelines for their own sector, assessing gaps and collecting

technical solutions for improved compliance testing. Test proposals are based on technical and

scientifically demonstrated justification. The guidelines will contain a common introduction and the

majority of association’s guidelines will be on their corresponding website.

4.3 Food consumption and packaging use data

At this point in time38, the default conventional assumption for packed food in Europe is that every

day an adult person consumes 1 kg of food packaged in a 1 dm3 cube with a surface area of 6 dm2.

It is assumed that the cube is covered by a single type of the same food contact material and the

food is the most aggressive extractor of the substance. The individual has the same exposure every

day throughout his life. This assumption is the basis for a default food contact rate for the material of

6 dm2/ person/ day.

This is a conservative assumption that does not reflect any real consumption pattern. In reality only

a certain percentage of the daily consumed food is packaged in any one food contact material and

within plastics a certain percentage is used to package aqueous food, acidic food, alcoholic food,

and fatty food.

EU food packaging cube

Total volume = 1 dm3

Total surface = 6 dm2

4.4 Derivation Estimated Daily Intake number – three steps approach

In order to calculate the estimated exposure, a tiered, three steps approach is suggested starting

with a simple worst case calculation up to very accurate estimates using highly sophisticated

probabilistic assessment models. Such refinements require additional information and are more

complex and resource intensive than the conservative approach.

STEP 1: WORST CASE EXPOSURE CALCULATION BASED ON EUROPEAN DEFAULT

ASSUMPTIONS

Based on the default assumption in Europe that every day an adult person consumes 1 kg of food

packaged in a 1 dm3 cube, an estimated worst-case daily intake number can be calculated using the

following simple formula.

38

This might change in light of the revision of EFSA’s note for guidelines, which is currently on-going.

10 cm

10 cm

10 cm

38

EDIworst case (mg/person/day) = 1 kg food/person/day * Migration (mg/kg food)

STEP 2: FRF CORRECTED EXPOSURE CALCULATION FOR LIPOPHILIC SUBSTANCES IN

FATTY FOOD39

At this time, to account for the fact that 95% of the population consumes less than 200 g of fat per

day, the Regulation facilitates dividing the migration levels of lipophilic substances into foods

containing more than 20% fat, with a Fat Reduction Factor (FRF). The FRF may vary from 1 to 5

(FRF=1 for food with a fat content of 20% and FRF=5 for food with fat content of 100%). The FRF

corrected exposure number can be calculated using the following formula.

1 kg food/person/day * Migration (mg/kg food)

EDIFRF corrected (mg/person/day) = -----------------------------------------------------------

FRFlipophilic substances

Annex V chapter 4.1 of the Regulation (EU) No. 10/2011 describes the application of the FRF and

the specific cases where the FRF cannot be applied (e.g. infant food).The application of the FRF

shall not lead to a specific migration exceeding the overall migration limit.

STEP 3 REFINED EXPOSURE CALCULATION USING FOOD DISTRIBUTION/CONSUMPTION

FACTORS

a) FDA exposure assessment

The US Food and Drug Administration (FDA) has generated consumption data for packaged

food which are used in risk assessments for regulatory purposes. The food consumption data for

different packaging materials (Consumption Factors and Food-Type Distribution Factors) are

accessible on the FDA website together with detailed guidance on how to calculate consumer

exposure to substances migrating from packaging materials40.

Under FDA approach, the term "Consumption Factor" (CF) is used to describe the fraction of the

daily diet expected to contact specific packaging materials. FDA assumes that an individual

consumes 3 kg of packaged food (1.5 kg solid and 1.5 kg liquid) per day of which about 80% is

packaged in plastics. The CF represents the fraction of daily consumed packaged food that is

packed in a certain packaging material. It somehow reflects the fact that only a certain part of the

daily food is packaged and especially packaged in a certain material. Table 1 of the a.m. FDA

guidance document shows specific packaging materials e.g. PET with a CF of 0.16. This means

39 http://www.tandfonline.com/doi/pdf/10.1080/02652030500157700 40

FDA guidance document for submission of a Food Contact Notification http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/IngredientsAdditivesGRASPackaging/ucm081818.htm#iie1a)

39

that 16 % of the daily diet (of 3 kg liquid and solid food) is packaged in PET. Or around 20 % of

the daily diet is packaged in metal cans (CF is 0.17 for coated and 0.03 for uncoated metal

containers). Since consumer behavior changes these CF-values need to be adapted from time to

time.

(http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/Ingre

dientsAdditivesGRASPackaging/ucm081818.htm#aivti);.

The Food-type distribution factors (fT) reflect for each packaging material the fractions of all food

contacting each material that is aqueous, acidic, alcoholic and fatty. These values take into

account the fact that certain packaging materials are used for specific foods of the daily diet. i.e.

Coated paper are rarely used (<5%) for foods which fall in the category acidic or alcoholic.

However, they are mostly used (95%) for foods of type aqueous or fatty (e.g meat).

For the calculation of the overall migration of a food contact substance from a specific material

the analytical results of migration testing with food simulants (<M>) are combined in that way

using Food-type distribution factors (fT) to reflect the real life use of that packaging material. In a

second step Exposure (EDI) is calculated from Migration (<M>) considering market share and

usage of that material (CF).

These data (consumption factors and food type distribution factors) are available on the FDA

website41

The Estimated Daily Intake (EDI) (mg/person/day) of a substance migrating from a

plastic packaging material into a specific type of food can be calculated using the

following formula (calculated for a 60 kg person and an intake of 3 kg of packaged

food per day):

EDIFDA (mg/person/day) = 3 kg food/person/day * CF * <M> (mg/kg food)

EDI: Estimated Daily Intake (mg/person/day)

CF: Consumption Factor for the particular plastic

<M>: Migration level of substance from the plastic into the food (mg/kg food)

In case specific migration levels are available for the different types of food then the formula can

be further refined as follows:

<M> = faqueous and acidic .(M 10% ethanol)+falcohol.(M 50% ethanol)+ffatty .(Mfatty)

f: Food-Type distribution factors for the particular plastic

M: Migration levels of the substance measured in different types of food simulants

The website of US FDA contains a database with cumulative estimated daily intakes (CEDIs)

and acceptable daily intakes (ADIs) for a large number of food contact substances7.

41

http://www.fda.gov/Food/default.htm (Documents UCM081825 and UCM081818)

40

4.5 European consumer exposure assessment tools

4.5.1 Risk Assessment of non-intentionally added substances (NIAS) using the MATRIX

Only limited food consumption data are publicly available for plastic packaging materials in Europe6f.

The assessment of NIAS apart from their occurrence also requires exposure data for the specific

plastics materials.

The Matrix Project was jointly initiated, financed and supported by Cefic-FCA, European Plastics

Converters (EuPC), Flexible Packaging Europe (FPE) and PlasticsEurope8. Within the project

generic levels of migration into food for respective packaging plastics materials were derived. Above

these levels every migrant should be identified and assessed, however, below which the

corresponding exposure is so minor that further assessment could be neglected. This level has been

defined “Level of Interest (LOI)”: it is linked to each packaging material and will be a function of the

exposure of consumers to this material. The calculation of the LOI follows similar conditions as

applied to non-listed substances used behind a functional barrier as described in the articles of

Regulation (EU) No 10/2011.

For specific packaging not often used in a country or specific food not consumed often, the average

real surface exposure can be very low compared to the standard 600 cm2 (i.e. Aluminium foil in Italy

the surface exposure is 0.06 dm2/person/day) can lead to a high Limit of Interest. However, it was

agreed by industry, that even if the calculated LOI was above 100, we would apply a maximum

conventional LOI of 100 µg/dm2, for negligible or low surface exposure (Surface < 0.1

dm2/person/day).

The Matrix Project derived country data sets for Germany, France, Italy, Spain and United Kingdom

with the respective packaging surface to which consumers are exposed per plastic material group

and per consumed food and the respective calculation of LOIs.

Plastics material groups can be assessed on a country base to define the level where identified

migrants need to be further risk assessed or not.

If NIAS assessments are addressed using the Matrix method the data and assessments become

part of the supporting documentation of the products investigated at the respective stage in the

Plastics value chain.

In general, the same methodology applied here for NIAS can be used for any non-listed substances.

41

MATRIX EXAMPLE

* Further information about MATRIX can be found under: http://matrixcalculation.eu/

PACKAGING USE DATA FROM THE EXPOSURE MATRIX PROJECT

The Exposure Matrix project collected information on the composition and amount of food packaging

in five EU countries (DE, UK, IT, FR, ES) in the period 2003-2006. The scope of the project can be

described as direct food contact packaging for pre-packed food sold to consumers in retail channels.

The packaging use data is a combination of data obtained from third party market survey

organisations and information collected from member companies of the participating associations.

More background information can be found here (registration and log-in required)42

The collected packaging use data has been combined with official dietary surveys from the five

countries by methods of stochastical modelling to result in a value for the average daily consumer

exposure to an area of food packaging, split up per food type and per packaging type. The

calculated values are an average (i.e. the 50th percentile of the distribution) because the intention

was to look at results over the entire diet or large parts of it and no person can be the extreme

consumer for every food. Nevertheless several factors of over-estimation remain (100% market

shares, 100% packaging loyalty, 100% packed foods, certain double-counting in the data, etc.) and

also the average value was seen to be higher than the median. Therefore the average packaging

area exposure value is seen to represent a sizeable majority of consumers.

42

https://matrixcalculation.eu/matrix/matrix.nsf/mv-exposure-matrix-14-01-2008.pdf

42

The results from the Exposure Matrix project can be used in certain cases to get a more refined

insight into the use of specific packaging materials than what is possible with the standard EU

convention that the diet consists of 1 kg food in contact with 6 dm² of each packaging material.

The following points need to be considered:

The data only covers the scope of the project as outlined above. If a substance is being

assessed for an application outside the scope, this cannot be done on the basis of this

Exposure Matrix project.

This methodology would likely give a false result if the substance or material is known to

have a disproportionally large contribution to consumer exposure from FCM applications not

covered by the scope. An example would be polystyrene, known to be heavily used in

vending and fast food outlets, or paper & board known to be heavily used in secondary

packaging and fast food. Exceptions may apply when assessing substances only used in

specific grades and/or specific applications which are known to be within the scope of the

Exposure Matrix project.

This method is not intended as an assessment for “all foods, all materials” as then the

numbers add up to being significantly more severe than the standard EU 6 dm²/kg scenario

(which is proof that there are still significant overestimations in the Exposure Matrix

methodology).

Refer to the Exposure Matrix results tables giving the average exposure (in dm²/person/day) for 51

food groups (50 for Germany) against 77 packaging materials (54 for Germany). In extracting

information from these tables, several different scenarios are possible depending on the origin of the

substance and the information available about the use of the substance or its parent material. It is

recommended to look at all five countries and select the highest result.

The following considerations can be made:

a) Do not look at the value of an individual matrix cell. These tables were not designed to be

used at this level of detail. First of all, they are an average value instead of a 95th percentile.

Also, there are inevitably data inaccuracies both in the underlying packaging data as well as

(and possibly even more) in the dietary survey data. These will only average out if a larger

consumption pattern is being looked at.

b) When the substance being assessed is known to be present in one or more specific

components (e.g. a polymer type), the relevant result is the value reported at the bottom of

the column (i.e. the sum over that entire column) in which that component is present.

1) If the component is present in more than one column, add up all relevant results.

2) Take into account that in certain material descriptions, certain components may be

present “anonymously”:

i. Example 1: acid/ester copolymer PE types are listed as an individual material

description, nevertheless they will also make up a certain percentage of any

other column that mentions the more generic term “PE” or “plastic”.

43

ii. Example 2: adhesives can be assumed to be present in all material

descriptions described as having two or more substrate layers (e.g. materials

such as “APET/PE”). Refinements based on type of adhesive chemistry,

would depend on expert judgement, to be documented.

iii. Example 3: the presence of inks or coatings cannot be easily derived from the

material descriptions listed, but may be allocated based on expert judgement.

Refinements based on type of ink or coating chemistry, would depend on

expert judgement, to be documented.

3) For niche products known and documented to have a very small market share, it is

allowed to take a fraction of the values reported, but not less than 10% [value to be

discussed].

c) The procedure described under point (b) may be modified by not making the sum over all

food groups, but only a sub-set of them, if there is documented justification for doing so:

When the substance is known not to migrate in all simulants, then the food groups

can be selected for which the relevant simulant(s) is/are applicable according to the

table in Annex III of Regulation 10/2011.

i. Example 1: A non-volatile substance can be considered not to migrate into

dry food (possibly to be confirmed by testing in Tenax), in which case the food

groups corresponding to dry food can be eliminated.

ii. Example 2: A lipophilic substance is expected to migrate only into fat/oil,

therefore the food groups for which simulants D1 or D2 are assigned, can be

selected for the evaluation.

iii. Example 3: An FCM substance marketed only for non-fatty food contact

applications, can be evaluated by eliminating food groups for which only

simulant D1 or D2 has been assigned.

When the substance has a very specific function related to a very specific food

packaging application, the food group(s) that cover this packaging application can be

selected – but taking into account the point (a) above.

I. Example: A substance known (but see point e) to be only used in can

coatings, would be only relevant for those food groups where cans are used,

and only for the column covering metal packaging (“other packaging” in the

German data; “other mono-substrate non-plastic” for the other countries).

d) When the substance being assessed is known to be associated to a specific FCM substance,

e.g. a NIAS resulting from an additive, the parent substance can be used in this methodology

if that makes it easier to collect information about the use and presence of the substance.

e) It can be difficult to know where in the market / in which kinds of materials a certain

substance is used or present. Expert judgement usually only extends to that specific market

segment the company is active in. Inventory lists may be an indication of the various uses of

certain substances.

44

The result of the exercise carried out along the previous points (a) through (e) is typically a single

value which is the average consumer exposure to the relevant materials, in dm²/person/day, for the

worst case country of the five countries included in the project. Knowing the migration into food or

food simulants allows a calculation of the concentration in the diet of the substance at hand. This

then constitutes an exposure assessment relevant for the risk assessment of the substance:

If the migration is not known, certain assumptions can be made based on concentration in

the material, typical worst case thickness of the material, and assuming complete transfer of

the substance.

If the migration level is known to be different in different simulants (which would typically be

the case) then the exercise under points (a) through (e) can be carried out by splitting the

food groups according to the simulant assignment. This results in double counting for those

food groups where more than one simulant is assigned – this is seen as an extra safety

margin.

4.5.2 FACET Project

Within the 7th Framework Research Programme, Europe has developed a new tool for exposure of

substances migrating from food contact packaging. FACET (Flavours, Additives and food Contact

material Exposure Task) is an EU-funded project aimed at estimating exposure to three types of food

chemicals: food additives, flavourings and migratable substances from food contact materials. The

FACET project which was officially finished in October 2012 developed a software tool that models

exposure to substances migrating from food contact material on a country base for the EU

population. The probabilistic exposure results are based on comprehensive pan-European food

consumption and food packaging data encrypted into the software.43

The FACET software allows to calculate the consumer exposure to direct food additives and

flavouring substances, as well as to migrants from food contact materials. In the area of FCM

migrants, distinction is made between metal packaging and non-metal packaging. In what follows,

only non-metal packaging is considered.

In order to calculate the exposure, the model combines the following information inputs:

food consumption, as recorded in national dietary surveys;

43

See http://ihcp.jrc.ec.europa.eu/our_activities/food-cons-prod/chemicals_in_food/FACET

45

structure, size and market share of the various FCM used for each food;

composition of each component used in the FCM structure – identity of substances,

concentration range, probability of presence in the FCM;

contact conditions (range of temperature and time) during conditioning, storage and use of

the packed food product.

The FACET model originally had the intention of recording industry-collected data covering the three

last bullet points in the list above. This information is embedded as encrypted data files in the

FACET software. The user only has to run a “pre-population” (once per country) which performs a

calculation of the migration levels into food, and can then proceed to run the “assessment” which is

the actual exposure calculation.

The big advantage of this approach is that it allows an exposure calculation across the consumer’s

complete diet (per country, per dietary survey) without requiring extensive knowledge by the user of

the use and composition of all FCM in all packaging sectors – a level of expertise considered

unrealistic for any user.

Experience has shown however that severe data gaps exist in the encrypted background data, so

that the results may be completely unreliable. In particular, the presence of certain substances in

polymers was not correctly recorded. Therefore the use of this pre-loaded data within the FACET

software is not recommended.

Two alternative methodologies exist within the FACET software to by-pass the unreliable

background data: the “new substance wizard” and the “new packaging wizard”.

In the “new substance wizard” the user has the option to create a “new substance” or a “NIAS” –

there is no fundamental difference between the two. The substance is defined by a name and its

molecular weight and log P (octanol/water partitioning). The option to “replace existing substance” or

“NIAS associated with existing substance” should not be chosen as this refers back to the unreliable

encrypted background data. Instead, the user should select the material(s) containing the substance

or NIAS and self-define the concentration and market share of the substance in each of the materials

selected. Running the new substance wizard results in the new substance appearing in the list of

substances to be selected (at the very end, as “NS-1, NS-2” etc.) in the pre-population wizard (to

calculate the migration). In doing this, the program uses the encrypted background data on use and

structure of the FCM, but by-passes the background data on composition of the individual

components in the FCM. After running the pre-population, the exposure assessment can be started.

In the “new packaging wizard” the user first defines a substance (again with name, molecular weight

and log P) or selects it from the listed existing substances, and then proceeds to define the structure

of the food contact material in which the substance is present. This requires layer thickness(es),

concentration, surface/volume ratio. The user also has to define the food, pack size and food contact

conditions.An advantage of running the new packaging wizard, is that it gives the user access to the

results of the migration calculation. This is under “My Migration Data” in the “My Data” tab. The “view

pack table” button gives the migration results in mg/kg food. It should be noted that this is a

probabilistic migration model and is not designed to over-estimate the migration level. It is not a

validated model that can be used in the context of SML compliance. Running the new packaging

wizard includes a migration calculation, so no separate pre-population run is needed before starting

46

the exposure assessment. When starting the exposure calculation, select “use my concentration

data” and find the appropriate entry under “my migration data”.

In summary, the new substance wizard (NSW) and the new packaging wizard (NPW) allow the user

to perform a targeted investigation but have as disadvantage that they require the user to have some

expert knowledge as input into the software:

NSW requires that the user find the molecular weight of the substance, and knows which

material(s) the substance is present in along with concentration and market share.

NPW requires that the user find the molecular weight of the substance, and knows which

material(s) the substance is present in along with concentration and market share, what the

structure is of the FCM in which the material in question is used, and what the food type and

food contact conditions are for each application.

The NSW can be used to calculate the exposure from a substance present in a particular material or

even several materials, integrated over all FCM structures in which those materials are present, and

integrated over all foods for which those FCM are used. The output of an “all foods, all materials”

exposure run could therefore be seen as an exposure in the diet. Some care is needed in the

interpretation of these results, as it is unlikely that the user has sufficient knowledge of the use of a

substance across all different packaging sectors.

The NPW is more suitable for the risk assessment of a single packaging application known to the

user. It takes much more effort (separate runs) to generate information across a range of foods, and

is therefore unlikely to ever approach an exposure result in the diet. The NPW gives access to the

migration calculation which is very useful in itself, but is not a validated model for SML compliance

purposes.

The NSW and NPW have been tested for metal packaging and for plastics and inks. They are likely

to work equally well for plastic-like materials such as adhesives. Other materials, most notably paper

& board, remain to be investigated.

As a final note, the user should be aware of the limitations of the scope of the FACET project. This

can perhaps best be described as “consumer retail”. This implies that other stages in retail e.g. B2B

trade of bulk foods or food ingredients, as well as food outlets such as vending, restaurants, fast

food outlets etc., or drinking water infrastructure, or home cooking and food storage, are not covered

at all.

47

5. RISK CHARACTERISATION OR RISK ASSESSMENT

In the final risk characterisation step, the typical exposure level to the substance in the daily diet (the

Estimated Daily Intake) is compared to the maximum tolerable exposure level (the Tolerable Daily

Intake). Alternatively the migration level into the food is compared to a self-derived Specific Migration

Limit (SML). As long as the Estimated Daily Intake is below the Tolerable Daily Intake or the

migration level under the typical condition of use is below the self-derived SML, the use of the

substance is considered safe.

Exposure (EDI) < exposure threshold (TDI or TTC) or Specific migration < self-derived

SML SAFE

If a safe use in the specific application cannot be immediately concluded, then either the risk

assessment could be refined (refine the exposure estimation or generate more toxicological data) or

the exposure to the substance (migration level) could be reduced.

The risk assessment must be reviewed regularly to take into account the evolution of the knowledge

relating to the toxicity of the substance and if the conditions of use are changed or different.

Exposure (EDI) < Exposure treshold (TDI or TTC)

Or Specific migration < self derived SML

Exposure (EDI) > Exposure treshold (TDI or TTC) Or

Specific migration > self derived SML

Product is

SAFE !

Product might

not be SAFE!

Refine risk assessment or risk reduction (e.g. reduce exposure)

48

REFERENCES

PlasticsEurope Guidance on Risk Assessment of non-listed substances (NLS) and non-intentionally

substances (NIAS° under Article 19 of Regulation 10/2011/ EU

Koster S, Boobis AR, Cubberley R, Hollnagel HM, Richling E, Wildemann T, Würtzen G, Galli CL.

Application of the TTC concept to unknown substances found in analysis of foods. Food Chem

Toxicol. 2011 Aug;49(8):1643-60.

Technical guidance safety assessments under REACH:

http://guidance.echa.europa.eu/guidance_en.htm

EFSA note for guidance for food contact materials:

http://ec.europa.eu/food/food/chemicalsafety/foodcontact/documents_en.htm

FDA guidance document - Toxicological assessment:

http://www.fda.gov/Food/default.htm (Document UCM081825)

FDA guidance document - Consumer Exposure assessment :

http://www.fda.gov/Food/default.htm (Document UCM081818)

ILSI publications (http://www.ilsi.org/Europe/Pages/HomePage.aspx)

- Threshold of Toxicological Concern (2005)

- The Acceptable Daily Intake: A Tool for Ensuring Food Safety (2000)

- Guidance for Exposure Assessment of Substances Migrating from Food Packaging Materials

(2007)

- Exposure from food contact materials (October 2002)

- Guidance for Exposure Assessment of Substances Migrating from Food Packaging Materials

(2007)

- Food Consumption and Packaging Usage Factors (1997)

- The Use of an Additional Safety Factor or Uncertainty Factor for Nature of Toxicity in the

Estimation of Acceptable Daily Intake and Tolerable Daily Intake Values

FDA CEDI/ ADI database

http://www.fda.gov/Food/IngredientsPackagingLabeling/PackagingFCS/CEDI/default.htm

Benford D, Bolger PM, Carthew P, Coulet M, DiNovi M, Leblanc JC, Renwick AG, Setzer W,

Schlatter J, Smith B, Slob W, Williams G, Wildemann T. Application of the Margin of Exposure

(MOE) approach to substances in food that are genotoxic and carcinogenic. Food Chem Toxicol.

2010 Jan; 48 Suppl 1:S2-24.

49

ECHA Guidance on information requirements and chemical safety assessment.

Chapter R.8: Characterisation of dose [concentration]-response for human health. May 2008.

http://echa.europa.eu/documents/10162/13632/information_requirements_r8_en.pdf

Opinion of the Scientific Committee on a request from EFSA related to a harmonized approach for

Risk Assessment of substances which are both Genotoxic and Carcinogenic. The EFSA Journal

(2005) 282, 1-31. http://www.efsa.europa.eu/en/scdocs/scdoc/282.htm

ECETOC report TR 085 - Recognition of, and Differentiation between, Adverse and Non-adverse

Effects in Toxicology Studies December 2002.

ECETOC report TR 099 - Toxicological Modes of Action: Relevance for Human Risk Assessment

July 2006.

IPCS, 2005. Chemical-Specific Adjustment Factors for Interspecies Differences And Human

Variability: Guidance Document For Use Of Data In Dose/Concentration–Response Assessment.

World Health Organisation, the International Labour Organisation and the United Nations

Environment Programme. Harmonisation Project Document No. 2.

ECHA Guidance on information requirements and chemical safety assessment. Chapter R.6:

QSARs and grouping of chemicals. May 2008.

http://echa.europa.eu/documents/10162/13632/information_requirements_r6_en.pdf

Kroes, R. et al., 2004. Structure-based thresholds of toxicological concern (TTC): guidance for

application to Substances present at low levels in the diet. Food and Chemical Toxicology 42: p 65-

83.

Kroes, R. et al., 2000. Thresholds of Toxicological Concern for chemical substances present in the

diet: A practical Tool for Assessing the Need for Toxicity Testing, Food and Chemical Toxicology

30: p 255-312.

Cramer GM, Ford RA, Hall RL. 1978. Estimation of toxic hazard--a decision tree approach. Food

Cosmet Toxicol. 16(3):255-76.

Munro IC, Ford RA, Kennepohl E, Sprenger JG. 1996. Correlation of structural class with no-

observed-effect levels: a proposal for establishing a threshold of concern. Food Chem Toxicol.

34(9):829-67.

50

ToxTree version 2.5.8 (http://toxtree.sourceforge.net/) and Patlewicz G, Jeliazkova N, Safford RJ,

Worth AP, Aleksiev B. An evaluation of the implementation of the Cramer classification scheme in

the ToxTree software. SAR QSAR Environ Res. 2008;19(5-6):495-524. PubMed PMID: 18853299.

Walker R, Kroes R. [The threshold of toxicological concern]. Vopr Pitan. 2002; 71(1):42-4. Review.

Russian. PubMed PMID: 12018155.

Barlow SM, Kozianowski G, Würtzen G, Schlatter J. Threshold of toxicological concern for chemical

substances present in the diet. Report of a workshop, 5-6 October 1999, Paris, France. Food Chem

Toxicol. 2001 Sep;39(9):893-905. PubMed PMID: 11498266.

Kroes R, Kozianowski G. Threshold of toxicological concern (TTC) in food safety assessment.

Toxicol Lett. 2002 Feb 28;127(1-3):43-6. Review. PubMed PMID: 12052639.

Renwick AG. Toxicology databases and the concept of thresholds of toxicological concern as used

by the JECFA for the safety evaluation of flavouring agents. Toxicol Lett. 2004 Apr 1;149(1-3):223-

34. Review. PubMed PMID: 15093268.

Kroes R, Kleiner J, Renwick A. The threshold of toxicological concern concept in risk assessment.

Toxicol Sci. 2005 Aug;86(2):226-30. Epub 2005 Apr 13. PubMed PMID: 15829616.

Renwick AG. Structure-based thresholds of toxicological concern--guidance for application to

substances present at low levels in the diet. Toxicol Appl Pharmacol. 2005 Sep 1;207(2 Suppl):585-

91. Review. PubMed PMID: 16019047.

Munro IC, Renwick AG, Danielewska-Nikiel B. The Threshold of Toxicological Concern (TTC) in risk

assessment. Toxicol Lett. 2008 Aug 15;180(2):151-6. Epub 2008 May 22. Review. PubMed PMID:

18573621.

Felter S, Lane RW, Latulippe ME, Llewellyn GC, Olin SS, Scimeca JA, Trautman TD. Refining the

threshold of toxicological concern (TTC) for risk prioritisation of trace chemicals in food. Food Chem

Toxicol. 2009 Sep;47(9):2236-45. Epub 2009 Jun 14. PubMed PMID: 19531369.

Union Guidelines on Regulation (EU) No 10/2011 on plastic materials and articles intended to come

into contact with food -http://ec.europa.eu/food/safety/docs/cs_fcm_plastic-guidance_201110_en.pdf

51

Chapter 3.1

http://apps.echa.europa.eu/registered/registered-sub.aspx

http://www.efsa.europa.eu/en/publications.htm

http://www.oecd.org/chemicalsafety/risk-assessment/theoecdqsartoolbox.htm

http://www.cefic-lri.org/lri-toolbox

https://eurl-ecvam.jrc.ec.europa.eu/laboratories-

research/predictive_toxicology/qsar_tools/toxtree

http://www.dguv.de/dguv/ifa/Gefahrstoffdatenbanken/GESTIS-Stoffdatenbank/index-2.jsp

http://chem.sis.nlm.nih.gov/chemidplus/

http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen?TOXLINE

http://webnet.oecd.org/hpv/ui/SponsoredSubstances.aspx

http://www.nicnas.gov.au/chemical-information

http://www.cir-safety.org/ingredients

https://fcm.wiv-isp.be/

https://pubchem.ncbi.nlm.nih.gov/

http://www.chemspider.com/

http://esis.jrc.ec.europa.eu/

Klimisch H., Andreae M. and Tillmann U. (1997) - A systematic approach for evaluating the

quality of experimental toxicological and ecotoxicological data. Regul Toxicol Pharm, 25, 1-5.

and Schneider K, Schwartz M, Burkholder I, Kopp-Schneider A, Edler L, Kinsner-Ovaskainen A,

Hartung T, Hoffmann S. (2009). “ToxR Tool”, a new tool to assess the reliability of toxicological

data. Toxicology Letters 189(2):138-144

https://eurl-ecvam.jrc.ec.europa.eu/about-ecvam/archive-publications/toxrtool

http://www.scirap.org/Page/Index/1e69f488-36b4-4660-9986-c529b9d9b40d/toxicity-framework

Beronius A, Molander L, Ruden C, Hanberg A (2014). Facilitating the use of non-standard in

vivo studies in health risk assessment of chemicals: A proposal to improve evaluation criteria

and reporting. Journal of Applied Toxicology 34(6): 607-617

Chapter 3.2 Uncertainty factors

Blackburn K, Stuard SB (2014). A Framework to Facilitate Consistent Characterisation of Read

Across Uncertainty. Regulatory Toxicology and Pharmacology 68: 353-362

ECETOC (2002). Report TR 085 - Recognition of, and Differentiation between, Adverse and

Non-adverse Effects in Toxicology Studies.

ECETOC report TR 099 - Toxicological Modes of Action: Relevance for Human Risk

Assessment July 2006.

52

ECHA (2012). Guidance on information requirements and chemical safety assessment Chapter

R.8: Characterisation of dose [concentration]-response for human health. Version 2.1. Available

at http://echa.europa.eu/documents/10162/13632/information_requirements_r8_en.pdf

IPCS (2005). Chemical-Specific Adjustment Factors for Interspecies Differences And Human

Variability: Guidance Document For Use Of Data In Dose/Concentration–Response

Assessment. World Health Organisation, the International Labour Organisation and the United

Nations Environment Programme. Harmonisation Project Document No. 2.

EFSA (2012). Guidance on selected default values to be used by the EFSA Scientific

Committee, Scientific Panels and Units in the absence of actual measured data. EFSA Journal

10(3):2579. Available at http://www.efsa.europa.eu/en/efsajournal/pub/2579

Chapter 3.3 The use of DNELs for Risk Assessment in Food Contact

Guidance on information requirements and chemical safety assessment

https://echa.europa.eu/documents/10162/13632/information_requirements_r8_en.pdf

REACh Registered Substances

https://www.echa.europa.eu/web/guest/information-on-chemicals/registered-substances

Guidance on Derivation of DNEL/DMEL from Human Data

https://echa.europa.eu/documents/10162/13632/r8_dnel_hd_draft_rev1-1_after_peg_en.pdf

Chapter 3.4 Read across

Benigni R, Bossa C (2008). Structure alerts for carcinogenicity, and the Salmonella assay

system: A novel insight through the chemical relational databases technology. Mutation

Research 659: 248-261

ECHA (2008). Guidance on information requirements and chemical safety assessment. Chapter

R.6: QSARs and grouping of chemicals. Available at

http://echa.europa.eu/documents/10162/13632/information_requirements_r6_en.pdf

Blackburn K, Bjerke D, Daston G, Felter S, Mahony C, Naciff J, Robison S, Wu S (2011). Case

studies to test: A framework for using structural, reactivity, metabolic and physicochemical

similarity to evaluate the suitability of analogs for SAR-based toxicological assessments.

Regulatory Toxicology and Pharmacology 60: 120-135

Blackburn K, Stuard SB (2014). A framework to facilitate consistent characterisation of read

across uncertainty. Regulatory Toxicology and Pharmacology 353-362

ECHA (2012). Practical Guide 5: How to report (Q)SARs. Version 2.0. Available at

http://echa.europa.eu/documents/10162/13655/pg_report_qsars_en.pdf

ECHA (2015). Grouping of Substances and Read across.

http://echa.europa.eu/support/grouping-of-substances-and-read-across

EFSA (2009). EFSA statement on the presence of 4-methylbenzophenone found in breakfast

cereals. The EFSA Journal (2009) RN-243, 1-19.

53

EFSA (2009). Scientific Opinion of EFSA prepared by the Panel on food contact materials,

enzymes, flavourings and processing aids (CEF) on Toxicological evaluation of benzophenone.

The EFSA Journal (2009) 1104, 1-30.

EFSA Scientific Committee (2012): Scientific Opinion on Exploring options for providing advice

about possible human health risks based on the concept of Threshold of Toxicological Concern

(TTC). EFSA Journal 10(7):2750 [103 pp.]. Available at

http://www.efsa.europa.eu/sites/default/files/scientific_output/files/main_documents/2750.pdf

ICH (International Conference on Harmonisation of Technical Requirements for Registration of

Pharmaceuticals for Human Use) (2014). Assessment and Control of DNA Reactive (Mutagenic)

Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk. ICH Harmonised Tripartite

Guideline M7. Public available at

http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Multidisciplinary/M7/M7

_Step_4.pdf

OECD (2007). Guidance Document on the Validation (Quantitative) Structure-Activity

Relationships ((Q)SAR) Models. OECD Environment Health and Safety Publications. Series on

Testing and Assessment No. 69. ENV/JM/MONO(2007)2. Public available at

http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2007)2&d

oclanguage=en

OECD (2014). Guidance on Grouping of Chemicals, Second Edition. Environment Health and

Safety Publications. Series on Testing and Assessment No. 194. ENV/JM/MONO(2014)4. Public

available under:

http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=ENV/JM/MONO(2014)4&doclanguage=en.

Patlewicz G, Ball N, Booggaard PJ, Becker RA, Hubesch B (2015). Building Scientific

Confidence in the Development and Evaluation of Read Across. Regulatory Toxicology and

Pharmacology 72: 117-133

Schilter B, Begnini R, Boobis A, Chiodini A, Cockburn A, Cronin MTD, Lo Piparo E, Modi S,

Thiel A, Worth A (2014). Establishing the Level of Safety Concern for Chemicals in Food without

the Need for Toxicity Testing. Regulatory Toxicology and Pharmacology 68:275-296

Schultz TW, Amcoff P, Bergren E, Gautier F, Klaric M, Knight DJ, Mahony C, Schwarz M, White

A, Cronin MTD (2015). A Strategy for Structuring and Reporting a Read Across Prediction of

Toxicity. Regulatory Toxicology and Pharmacology 72: 586-601

Wu S, Blackburn K, Amburgey J, Jaworska J, Federle T (2010). A framework for using

structural, reactivity, metabolic and physicochemical similarity to evaluate the suitability of

analogs for SAR-based toxicological assessments. Regulatory Toxicology and Pharmacology

56: 67-81

Chapter 3.5 TTC

EFSA event report (2016). “Review of the Threshold of Toxicological Concern (TTC) approach

and development of new TTC decision tree.” EFSA Supporting publication 2016:EN-1006.

Kroes R, Galli CL, Munro I, Schilter B, Tran L-A, Walker R, Würtzen G (2000) Threshold of

toxicological concern for chemical substances present in the diet: a practical tool for assessing

the need for toxicity testing. Food and Chemical Toxicology 38, s. 255–312.

54

Kroes R, Kleiner J, Renwick A (2005) The threshold of toxicological concern concept in risk

assessment Toxicological Science 86 (2), s. 226–230.

Kroes R, Kozianowski G (2002) Threshold of toxicological concern (TTC) in food safety

assessment. Toxicology Letters 127, s. 43–46.

Munro IC, Ford RA, Kennepohl E, Sprenger JG (1996) Correlation of structural class with no-

observed-effect levels: A proposal for establishing a threshold of concern. Food and Chemical

Toxicology 34, s. 829–867.

Kroes R, Renwick A, Cheesman M, Kleiner J, Mangelsdorf I, Piersma A, Schilter B, Schlatter J,

van Schothorst F, Vos JG, Wurtzen G (2004) Structured-based thresholds of toxicological

concern (TTC): Guidance for application to substances present at low levels in the diet. Food

and Chemical Toxicology 42, s. 65–83.

Hennes EC (2012) An overview of values for the threshold of toxicological concern. Toxicology

Letters 20;211(3), s. 296-303.

Cheeseman MA, Machuga EJ, Bailey AB (1999) A tiered approach to threshold of regulation.

Food and Chemical Toxicology 37, s. 387–412.

Cramer GM, Ford RA, Hall RL (1978) Estimation of toxic hazard – a decision tree approach.

Food and Chemical Toxicology 16, s. 255–276.

EFSA (2012). “Scientific Opinion on exploring options for providing advice about possible human

health risks based on the concept of Threshold of Toxicological Concern (TTC).” EFSA Journal

2012;10(7):2750.

Koster, S. et al. (2011). “Application of the TTC concept to unknown substances found in

analysis of foods.” Food Chem Toxicol 49:1643-60.

Chapter 3.7 Nano

EFSA Scientific opinion “Guidance on the risk assessment of the application of nanoscience and

nanotechnologies in the food and feed chain” 2011

EFSA Scientific opinion “Recent Development in the risk assessment of chemicals in food and

their potential impact on the safety assessment of substances used in food contact materials”

2015

EFSA Scientific “Opinion on the safety evaluation of the substance, titanium nitride,

nanoparticles,for use in food contact materials” 2012

Chapter 3.8 Endocrine Disruptors

EFSA (2013). Scientific Opinion on the hazard assessment of endocrine disruptors: Scientific

criteria for identification of endocrine disruptors and appropriateness of existing test methods for

assessing effects mediated by these substances on human health and the environment. EFSA

Journal 11(3):3132

IPCS (2002). Global assessment of the state-of-the-science of endocrine disruptors. Geneva,

Switzerland, World Health Organization, International Programme on Chemical Safety.

55

ECETOC (2009). Guidance on Identifying Endocrine Disrupting Effects. Technical Report No.

106.

Chapter 3.9 Bioaccumulation / bioavailability in the human body

Regulation (EU) N° 10/2011on Plastic Materials and Articles Intended to Come into contact with

Food

EFSA Draft Scientific Opinion on Recent Developments in the Risk Assessment of Chemicals in

Food and their Potential Impact on the Safety Assessment of Substances Used in Food Contact

Materials

ILSI Guidance on Best Practices on the Risk Assessment of Non Intentionally Added

Substances (NIAS) in Food Contact Materials and Articles

56

Food Contact Additives - Sector Group

The European Chemical Industry Council - Cefic

Av. E. van Nieuwenhuyse 4 (box 2) -B - 1160 Brussels fca@cefic.be // http://fca.cefic.org/